Project

General

Profile

action #73633

OSD partially unresponsive, triggering 500 responses, spotty response visible in monitoring panels but no alert triggered (yet)

Added by okurz about 1 year ago. Updated 11 months ago.

Status:
Resolved
Priority:
Urgent
Assignee:
Target version:
Start date:
2020-10-20
Due date:
2020-11-17
% Done:

0%

Estimated time:

Description

Observation

https://monitor.qa.suse.de/d/WebuiDb/webui-summary?orgId=1&from=1603190156643&to=1603196975018
shows that at around 2020-10-20 12:39 the HTTP response time from osd increased and users reported spotty connection and 500 responses "unresponsive" during that time, e.g. in https://chat.suse.de/channel/testing?msg=aix9KNXwkWowTd7FA . The spotty response is visible in our monitoring panels we no alert triggered so far in grafana because we do not want the unspecific "No Data" alerts.

Cause, solution and test

ip6tables-save.firewalld.txt (5.98 KB) ip6tables-save.firewalld.txt Dump without default route over v6 nicksinger, 2020-11-04 15:00
ip6tables-save.susefirewall.txt (3.73 KB) ip6tables-save.susefirewall.txt Dump with default route over v6 nicksinger, 2020-11-04 15:00

Related issues

Related to openQA Infrastructure - action #75016: [osd-admins][alert] Failed systemd services alert (workers): os-autoinst-openvswitch.service (and var-lib-openqa-share.mount) on openqaworker-arm-2 and othersResolved2020-10-21

Related to openQA Infrastructure - action #75055: grenache-1 can't connect to webui's over IPv4 onlyResolved2020-10-22

Related to openQA Infrastructure - action #76828: big job queue for ppc as powerqaworker-qam-1.qa and malbec.arch and qa-power8-5-kvm were not activeResolved2020-10-31

Related to openQA Infrastructure - action #68095: Migrate osd workers from SuSEfirewall2 to firewalldResolved2020-06-15

Related to openQA Infrastructure - action #80128: openqaworker-arm-2 fails to download from openqaResolved2020-11-21

Has duplicate openQA Infrastructure - action #77995: worker instances on grenache-1 seem to fail (sometimes?) to connect to web-uis Rejected2020-11-16

Copied to openQA Infrastructure - action #78127: follow-up to #73633 - lessons learned and suggestionsResolved

History

#1 Updated by okurz about 1 year ago

  • Status changed from New to In Progress
  • Assignee set to okurz

coolo looked into the issue again this morning, coolo stating "we have a whopping 173 apache slots getting an artefact upload atm, and according to strace they get uploaded in bytes not MBs, SLES-15-SP3-s390x-63.1@s390x-kvm-sle15-minimal_with_sdk63.1_installed_withhome.qcow2: Processing chunk 520/3037, avg. speed ~19.148 KiB/s […] the workers have been restarted by salt now - but I stopped the scheduler and so far the 44 jobs running seem to run fine , https://stats.openqa-monitor.qa.suse.de/d/WebuiDb/webui-summary?orgId=1&from=1603183682771&to=1603199316026 - so the problem started around the time that Nick changed IP routes yesterday, not saying what's cause and what is symptom - but they are surely related […] So somehow suddenly all workers decided to slowdown uploads 🙂 […] So it seems to work all fine again - and all I did was turning it off and on again 😞".

since the last problematic incident we have https://stats.openqa-monitor.qa.suse.de/d/WebuiDb/webui-summary?orgId=1&refresh=30s&fullscreen&panelId=2&from=now-2d&to=now and I don't see anything severe showing up there at least. so likely something different? Although I can see that the number of database connections looks different at least since 2020-10-20 12:00
The apache response times in
https://stats.openqa-monitor.qa.suse.de/d/WebuiDb/webui-summary?tab=alert&orgId=1&from=now-2d&to=now&fullscreen&panelId=84&edit
show a significant increase in the response time which we can alert on.

EDIT: Created https://gitlab.suse.de/openqa/salt-states-openqa/-/merge_requests/384 for improved monitoring based on apache response time. In the meantime we restarted the openqa-webui service multiple times as well as apache2 and nicksinger removed the manually added IPv6 routes from all machines except grenache-1.

#2 Updated by nicksinger about 1 year ago

As the problem really escalated yesterday after I enabled a manual IPv6 route and most of OSD's connections where over v6:

openqa:~ # ss -tpn4 | wc -l
55
openqa:~ # ss -tpn6 | wc -l
1585

I now removed this route from all workers again. The command I used for this was:

salt -l error -C 'G@roles:worker' cmd.run 'ip -6 r d default via fe80::1 dev $(ip r s | grep default | sed -n "s/^.*dev \(.*\) proto dhcp/\1/p")'

If we see other problems we can think about disabling IPv6 completely for now on the externally connected interfaces like this:

salt -l error -C 'G@roles:worker' cmd.run 'echo 1 > /proc/sys/net/ipv6/conf/$(ip r s | grep default | sed -n "s/^.*dev \(.*\) proto dhcp/\1/p" | xargs)/disable_ipv6'

#3 Updated by nicksinger about 1 year ago

We have the initial infra ticket from yesterday about the missing v6 route: https://infra.nue.suse.com/SelfService/Display.html?id=178626. In the meantime I stated that all our machines are affected and that we can see severe performance issues over v6. Might be worth to create a new/more explicit one once we're sure we can blame the network.

#4 Updated by okurz about 1 year ago

From https://stats.openqa-monitor.qa.suse.de/d/WebuiDb/webui-summary?tab=alert&orgId=1&from=now-3h&to=now&fullscreen&panelId=84&edit I don't see severe problems right now. I planned to start openqa-scheduler again at 0930Z unless I hear objections.

EDIT:

<nicksinger> any objections on disabling v6 on grenache completely? I want to see if it works better then yesterday with a missing route
<okurz> I suggest we only apply changes one at a time. Do you see severe problems with grenache-1 right now? I consider it the most important issue that openqa-scheduler is not running so no new jobs will be started
<okurz> started [openqa-scheduler service], btw I hope you guys can all see the annotations in https://stats.openqa-monitor.qa.suse.de/d/WebuiDb/webui-summary?tab=alert&orgId=1&from=now-3h&to=now&fullscreen&panelId=84&edit ? started, openqa-scheduler on osd again, monitoring the mentioned grafana panel. Updated https://progress.opensuse.org/issues/73633 and also commented in https://infra.nue.suse.com/SelfService/Display.html?id=178626 . Thanks Nick Singer for the ticket update and the EngInfra ticket reference and making sure that they understand the gravity 🙂

an alert for "apache response time" is deployed now and it's currently green.
I put the threshold on 500ms avg as I saw that the avg would creep up slowly so I think 1s could give us an alert a bit sooner but still not trigger falsely.

#5 Updated by okurz about 1 year ago

  • Related to action #75016: [osd-admins][alert] Failed systemd services alert (workers): os-autoinst-openvswitch.service (and var-lib-openqa-share.mount) on openqaworker-arm-2 and others added

#6 Updated by okurz about 1 year ago

  • Due date set to 2020-10-23
  • Status changed from In Progress to Feedback

For the past hours I was looking into #75016 which I assume to be related. Also I was monitoring grafana alerts (no new alerts during this time) and found no further problems. I am not aware of any current things that do not work. We can try changes regarding "IPv6" again maybe tomorrow as long as no new issues came up or the situation regressed.

#7 Updated by nicksinger about 1 year ago

  • Related to action #75055: grenache-1 can't connect to webui's over IPv4 only added

#8 Updated by okurz about 1 year ago

  • Due date changed from 2020-10-23 to 2020-10-24

osd itself seems to be fine but some machines have problems and are not conducting tests at all. Right now all three arm machines are not conducting tests. On openqaworker-arm-1 which was automatically rebooted (after crash) 5h ago all worker services fail to to reach osd as they try over IPv6 but fail due to the missing route.

What I did now:

echo /proc/sys/net/ipv6/conf/$(ip r s | grep default | sed -n "s/^.*dev \(.*\) proto dhcp/\1/p" | xargs)/disable_ipv6
systemctl restart openqa-worker@\* openqa-worker-cacheservice openqa-worker-cacheservice-minion.service os-autoinst-openvswitch.service

and tests start again but this is not persistent.

I guess we could call

salt -l error -C 'G@roles:worker' cmd.run 'echo net.ipv6.conf.all.disable_ipv6 = 1 > /etc/sysctl.d/poo73633_debugging.conf && sysctl --load /etc/sysctl.d/poo73633_debugging.conf && systemctl restart openqa-worker@\* openqa-worker-cacheservice openqa-worker-cacheservice-minion.service os-autoinst-openvswitch.service'

I called that for openqaworker-arm-1 and openqaworker-arm-2 now only. qa-power8-5.qa.suse.de was not reachable and also IPMI SoL gave me nothing so I called power reset and after the machine is up also here like in #75016 the mount point service var-lib-openqa-share.mount failed and I fixed that by restarting with systemctl restart var-lib-openqa-share.mount. I did not remove IPv6 or anything, tests started up but not sure if they will work fine. I can't reach malbec.arch neither ssh nor over IPMI so no progress there.

EDIT: 2020-10-22 21:53: Retrying multiple times I can reach malbec.arch over ipmitool to confirm that "Chassis Power is on" but I can't get it to show anything on SoL so I can only try to trigger a power reset but running something like while [ $? != 0 ]; do ipmitool -4 -I lanplus -H fsp1-malbec.arch.suse.de -P $pass power reset && break; done for about 30m on both my computer as well as login1.suse.de fail to establish a session.

EDIT: 2020-10-22 23:40: At a later time I managed to "get through" to malbec and could trigger a power reset. It is conducting tests fine again right now.

EDIT: 2020-10-26 09:24: Applied the same ipv6 disablement from above to grenache-1.qa which failed to run any tests.

#9 Updated by nicksinger 12 months ago

So I dug a little more ending up hijacking openqaworker3 as my debugging host. First of, I installed tcpdump to be capable of wireshark tracing over ssh. Nothing too unexpected there besides router advertisements completely missing on the interface for the machine itself. I was however able to spot "Router Solicitation" originating from a QEMU mac (which should only happen if there was a previous RA, so the SUTs can see the router?). I continued probing for all routers (ping ff02::2 - ff02::2 is the multicast address for all routers):

64 bytes from fe80::56ab:3aff:fe16:ddc4%eth0: icmp_seq=1 ttl=64 time=0.067 ms
64 bytes from fe80::56ab:3aff:fe16:dd73%br0: icmp_seq=1 ttl=64 time=0.391 ms (DUP!)
64 bytes from fe80::56ab:3aff:fe24:358d%br0: icmp_seq=1 ttl=64 time=0.407 ms (DUP!)
64 bytes from fe80::2e60:cff:fe73:2ac%br0: icmp_seq=1 ttl=64 time=0.422 ms (DUP!)
64 bytes from fe80::ec4:7aff:fe7a:7896%br0: icmp_seq=1 ttl=64 time=0.471 ms (DUP!)
64 bytes from fe80::ec4:7aff:fe99:dcd9%br0: icmp_seq=1 ttl=64 time=0.486 ms (DUP!)
64 bytes from fe80::ec4:7aff:fe43:d6a8%br0: icmp_seq=1 ttl=64 time=0.484 ms (DUP!)
64 bytes from fe80::fab1:56ff:fed2:7fcf%br0: icmp_seq=1 ttl=64 time=0.500 ms (DUP!)
64 bytes from fe80::56bf:64ff:fea4:2315%br0: icmp_seq=1 ttl=64 time=0.530 ms (DUP!)
64 bytes from fe80::6600:6aff:fe73:c434%br0: icmp_seq=1 ttl=64 time=0.529 ms (DUP!)
64 bytes from fe80::529a:4cff:fe4c:e46d%br0: icmp_seq=1 ttl=64 time=0.554 ms (DUP!)
64 bytes from fe80::1a03:73ff:fed5:6477%br0: icmp_seq=1 ttl=64 time=0.560 ms (DUP!)
64 bytes from fe80::9a90:96ff:fea0:fc9b%br0: icmp_seq=1 ttl=64 time=0.569 ms (DUP!)
64 bytes from fe80::200:5aff:fe9c:4a11%br0: icmp_seq=1 ttl=64 time=0.567 ms (DUP!)
64 bytes from fe80::3d57:e68f:6817:810f%br0: icmp_seq=1 ttl=64 time=0.579 ms (DUP!)
64 bytes from fe80::ec4:7aff:fe7a:789e%br0: icmp_seq=1 ttl=64 time=0.587 ms (DUP!)
64 bytes from fe80::fab1:56ff:febe:b857%br0: icmp_seq=1 ttl=64 time=0.585 ms (DUP!)
64 bytes from fe80::1a66:daff:fe32:4eec%br0: icmp_seq=1 ttl=64 time=0.602 ms (DUP!)
64 bytes from fe80::1a66:daff:fe31:9434%br0: icmp_seq=1 ttl=64 time=0.627 ms (DUP!)
64 bytes from fe80::862b:2bff:fea1:28c%br0: icmp_seq=1 ttl=64 time=0.651 ms (DUP!)
64 bytes from fe80::b002:7eff:fe38:2d23%br0: icmp_seq=1 ttl=64 time=0.660 ms (DUP!)
64 bytes from fe80::d8a9:36ff:fe86:98b7%br0: icmp_seq=1 ttl=64 time=0.676 ms (DUP!)
64 bytes from fe80::3617:ebff:fe9e:6902%br0: icmp_seq=1 ttl=64 time=0.757 ms (DUP!)
64 bytes from fe80::fab1:56ff:feb8:367e%br0: icmp_seq=1 ttl=64 time=1.02 ms (DUP!)
64 bytes from fe80::2de:fbff:fee3:dafc%br0: icmp_seq=1 ttl=64 time=1.24 ms (DUP!)
64 bytes from fe80::2de:fbff:fee3:d77c%br0: icmp_seq=1 ttl=64 time=2.84 ms (DUP!)

It is very interesting to see so many entries in here. I still need to figure out how exactly how to read this but basically you can see that only one response came from eth0 while all the others came from our bridge on worker3. If all the br0 answers are actually from SUTs is yet unclear to me. But it could show a first problem.

I also fount the following which I just leave here for me to parse later:

openqa:~ # salt -l error -C 'G@roles:worker' cmd.run 'ip -6 neigh'
openqaworker8.suse.de:
openqaworker3.suse.de:
    fe80::1 dev br0 lladdr 00:00:5e:00:02:02 router STALE
openqaworker9.suse.de:
    fe80::a3c9:d83f:17aa:8999 dev eth1 lladdr d4:81:d7:5a:a3:9c STALE
    fe80::36ac:19a7:3193:7081 dev eth1 lladdr 0a:00:00:00:00:33 STALE
    fe80::216:3eff:fe48:17ff dev eth1 lladdr 00:16:3e:48:17:ff STALE
    fe80::5054:ff:fe44:d766 dev eth1 lladdr 52:54:00:44:d7:66 STALE
    fe80::5054:ff:fe44:d765 dev eth1 lladdr 52:54:00:44:d7:65 STALE
    fe80::5054:ff:fe44:d768 dev eth1 lladdr 52:54:00:44:d7:68 STALE
    fe80::5054:ff:fe44:d767 dev eth1 lladdr 52:54:00:44:d7:67 STALE
    fe80::6600:6aff:fe75:72 dev eth1 lladdr 64:00:6a:75:00:72 STALE
    fe80::c3ab:62d0:2723:6249 dev eth1 lladdr 64:00:6a:75:00:72 STALE
    2620:113:80c0:8080::4 dev eth1  FAILED
    fe80::501:abb4:eb5c:6686 dev eth1 lladdr e4:b9:7a:e4:aa:ad STALE
    fe80::5054:ff:fe30:a4d9 dev eth1 lladdr 52:54:00:30:a4:d9 STALE
    fe80::208:2ff:feed:8f15 dev eth1 lladdr 00:08:02:ed:8f:15 STALE
    fe80::2af1:eff:fe41:cef3 dev eth1 lladdr 28:f1:0e:41:ce:f3 STALE
    fe80::1 dev eth1 lladdr 00:00:5e:00:02:02 router STALE
    fe80::ec4:7aff:fe7a:7736 dev eth1 lladdr 0c:c4:7a:7a:77:36 STALE
    fe80::4950:d671:f08c:c9c3 dev eth1 lladdr 18:db:f2:46:1e:1d STALE
    fe80::9249:faff:fe06:82d8 dev eth1 lladdr 90:49:fa:06:82:d8 STALE
    fe80::2de:fbff:fee3:d77c dev eth1 lladdr 00:de:fb:e3:d7:7c router STALE
    fe80::d681:d7ff:fe5a:a39c dev eth1 lladdr d4:81:d7:5a:a3:9c STALE
    fe80::800:ff:fe00:15 dev eth1 lladdr 0a:00:00:00:00:15 STALE
    fe80::2de:fbff:fee3:dafc dev eth1 lladdr 00:de:fb:e3:da:fc router STALE
    fe80::56ab:3aff:fe16:ddc4 dev eth1 lladdr 54:ab:3a:16:dd:c4 router STALE
    fe80::5054:ff:fe29:137f dev eth1 lladdr 52:54:00:29:13:7f STALE
    fe80::1a66:daff:fe00:bbaa dev eth1 lladdr 18:66:da:00:bb:aa STALE
    fe80::800:ff:fe00:32 dev eth1 lladdr 0a:00:00:00:00:32 STALE
    fe80::5054:ff:fef4:ecb8 dev eth1 lladdr 52:54:00:f4:ec:b8 STALE
    fe80::5054:ff:fe87:8cc4 dev eth1 lladdr 52:54:00:87:8c:c4 STALE
openqaworker6.suse.de:
    fe80::5054:ff:fe44:d767 dev eth0 lladdr 52:54:00:44:d7:67 STALE
    fe80::1 dev eth0 lladdr 00:00:5e:00:02:02 router STALE
    fe80::2de:fbff:fee3:dafc dev eth0 lladdr 00:de:fb:e3:da:fc router STALE
    fe80::1a66:daff:fe00:bbaa dev eth0 lladdr 18:66:da:00:bb:aa STALE
    fe80::800:ff:fe00:15 dev eth0 lladdr 0a:00:00:00:00:15 STALE
    fe80::9249:faff:fe06:82d8 dev eth0 lladdr 90:49:fa:06:82:d8 STALE
    fe80::56ab:3aff:fe16:ddc4 dev eth0 lladdr 54:ab:3a:16:dd:c4 router STALE
    fe80::5054:ff:fe44:d765 dev eth0 lladdr 52:54:00:44:d7:65 STALE
    fe80::d681:d7ff:fe5a:a39c dev eth0 lladdr d4:81:d7:5a:a3:9c STALE
    fe80::216:3eff:fe48:17ff dev eth0 lladdr 00:16:3e:48:17:ff STALE
    fe80::208:2ff:feed:8f15 dev eth0 lladdr 00:08:02:ed:8f:15 STALE
    fe80::800:ff:fe00:32 dev eth0 lladdr 0a:00:00:00:00:32 STALE
    fe80::6600:6aff:fe75:72 dev eth0 lladdr 64:00:6a:75:00:72 STALE
    fe80::5054:ff:fe30:a4d9 dev eth0 lladdr 52:54:00:30:a4:d9 STALE
    fe80::5054:ff:fe44:d768 dev eth0 lladdr 52:54:00:44:d7:68 STALE
    fe80::36ac:19a7:3193:7081 dev eth0 lladdr 0a:00:00:00:00:33 STALE
    fe80::ec4:7aff:fe7a:7736 dev eth0 lladdr 0c:c4:7a:7a:77:36 STALE
    fe80::2908:884f:5368:dda dev eth0 lladdr c8:f7:50:40:f4:69 STALE
    fe80::2af1:eff:fe41:cef3 dev eth0 lladdr 28:f1:0e:41:ce:f3 STALE
    fe80::5054:ff:fe87:8cc4 dev eth0 lladdr 52:54:00:87:8c:c4 STALE
    fe80::5054:ff:fe44:d766 dev eth0 lladdr 52:54:00:44:d7:66 STALE
    fe80::5054:ff:feb1:4de dev eth0 lladdr 52:54:00:b1:04:de STALE
    fe80::501:abb4:eb5c:6686 dev eth0 lladdr e4:b9:7a:e4:aa:ad STALE
    fe80::2de:fbff:fee3:d77c dev eth0 lladdr 00:de:fb:e3:d7:7c router STALE
    fe80::a3c9:d83f:17aa:8999 dev eth0 lladdr d4:81:d7:5a:a3:9c STALE
    fe80::5054:ff:fef4:ecb8 dev eth0 lladdr 52:54:00:f4:ec:b8 STALE
    fe80::4950:d671:f08c:c9c3 dev eth0 lladdr 18:db:f2:46:1e:1d STALE
    fe80::c3ab:62d0:2723:6249 dev eth0 lladdr 64:00:6a:75:00:72 STALE
    fe80::5054:ff:fe29:137f dev eth0 lladdr 52:54:00:29:13:7f STALE
QA-Power8-4-kvm.qa.suse.de:
    fe80::1 dev eth3 lladdr 00:00:5e:00:02:04 router STALE
    fe80::f46b:41ff:feb7:9502 dev eth3 lladdr f6:6b:41:b7:95:02 STALE
    fe80::2de:fbff:fee3:dafc dev eth3 lladdr 00:de:fb:e3:da:fc router STALE
    fe80::215:5dff:fe43:a241 dev eth3 lladdr 00:15:5d:43:a2:41 STALE
    fe80::2de:fbff:fee3:d77c dev eth3 lladdr 00:de:fb:e3:d7:7c router STALE
    fe80::f46b:44ff:fe50:f502 dev eth3 lladdr f6:6b:44:50:f5:02 STALE
    fe80::5054:ff:fe47:10e4 dev eth3 lladdr 52:54:00:47:10:e4 STALE
    fe80::216:3eff:fe32:3671 dev eth3 lladdr 00:16:3e:32:36:71 STALE
    fe80::216:3eff:fec3:d305 dev eth3 lladdr 00:16:3e:c3:d3:05 STALE
    fe80::f46b:45ff:fe75:7e02 dev eth3 lladdr f6:6b:45:75:7e:02 STALE
    fe80::ae1f:6bff:fe01:130 dev eth3 lladdr ac:1f:6b:01:01:30 STALE
    fe80::f46b:47ff:fe57:de02 dev eth3 lladdr f6:6b:47:57:de:02 STALE
    fe80::215:5dff:fe43:a23d dev eth3 lladdr 00:15:5d:43:a2:3d STALE
    fe80::dc86:c1ff:fe33:d97f dev eth3 lladdr de:86:c1:33:d9:7f STALE
    fe80::1e1b:dff:feef:735c dev eth3 lladdr 1c:1b:0d:ef:73:5c STALE
    fe80::216:3eff:fe32:6543 dev eth3 lladdr 00:16:3e:32:65:43 STALE
    fe80::e2d5:5eff:fea7:e824 dev eth3 lladdr e0:d5:5e:a7:e8:24 STALE
    fe80::215:5dff:fe43:a23b dev eth3 lladdr 00:15:5d:43:a2:3b STALE
    fe80::f46b:4aff:fef5:d602 dev eth3 lladdr f6:6b:4a:f5:d6:02 STALE
    fe80::215:5dff:fe43:a239 dev eth3 lladdr 00:15:5d:43:a2:39 STALE
    fe80::f46b:46ff:fe0a:3202 dev eth3 lladdr f6:6b:46:0a:32:02 STALE
    fe80::216:3eff:fe32:8923 dev eth3 lladdr 00:16:3e:32:89:23 STALE
    fe80::f46b:4fff:fe78:3902 dev eth3 lladdr f6:6b:4f:78:39:02 STALE
    fe80::5054:ff:fea2:abb2 dev eth3 lladdr 52:54:00:a2:ab:b2 STALE
    fe80::20c:29ff:fe20:339f dev eth3 lladdr 00:0c:29:20:33:9f STALE
    fe80::225:90ff:fe9a:cb5e dev eth3 lladdr 00:25:90:9a:cb:5e STALE
    fe80::423:f5ff:fe3c:2c73 dev eth3 lladdr 06:23:f5:3c:2c:73 STALE
    fe80::f46b:43ff:fed5:9d02 dev eth3 lladdr f6:6b:43:d5:9d:02 STALE
    fe80::ff:fee1:a5b4 dev eth3 lladdr 02:00:00:e1:a5:b4 STALE
    fe80::5054:ff:fe40:4a1e dev eth3 lladdr 52:54:00:40:4a:1e STALE
    fe80::ec4:7aff:fe6c:400a dev eth3 lladdr 0c:c4:7a:6c:40:0a STALE
    fe80::215:5dff:fe43:a23e dev eth3 lladdr 00:15:5d:43:a2:3e STALE
    fe80::f46b:45ff:fee9:d803 dev eth3 lladdr f6:6b:45:e9:d8:03 STALE
    fe80::215:5dff:fe43:a23c dev eth3 lladdr 00:15:5d:43:a2:3c STALE
    fe80::ff:fee0:a4b3 dev eth3 lladdr 02:00:00:e0:a4:b3 STALE
    fe80::5054:ff:fe55:613f dev eth3 lladdr 52:54:00:55:61:3f STALE
    fe80::20c:29ff:fe9d:6297 dev eth3 lladdr 00:0c:29:9d:62:97 STALE
openqaworker2.suse.de:
    fe80::5054:ff:fe30:a4d9 dev br0 lladdr 52:54:00:30:a4:d9 STALE
    fe80::4950:d671:f08c:c9c3 dev br0 lladdr 18:db:f2:46:1e:1d STALE
    fe80::2de:fbff:fee3:dafc dev br0 lladdr 00:de:fb:e3:da:fc router STALE
    fe80::ec4:7aff:fe7a:7736 dev br0 lladdr 0c:c4:7a:7a:77:36 STALE
    fe80::6600:6aff:fe75:72 dev br0 lladdr 64:00:6a:75:00:72 STALE
    fe80::5054:ff:fe29:137f dev br0 lladdr 52:54:00:29:13:7f STALE
    fe80::800:ff:fe00:15 dev br0 lladdr 0a:00:00:00:00:15 STALE
    fe80::56ab:3aff:fe16:ddc4 dev br0 lladdr 54:ab:3a:16:dd:c4 router STALE
    fe80::9249:faff:fe06:82d8 dev br0 lladdr 90:49:fa:06:82:d8 STALE
    fe80::1 dev br0 lladdr 00:00:5e:00:02:02 router STALE
    fe80::2af1:eff:fe41:cef3 dev br0 lladdr 28:f1:0e:41:ce:f3 STALE
    2620:113:80c0:8080::5 dev br0  FAILED
    fe80::a3c9:d83f:17aa:8999 dev br0 lladdr d4:81:d7:5a:a3:9c STALE
    fe80::d681:d7ff:fe5a:a39c dev br0 lladdr d4:81:d7:5a:a3:9c STALE
    2620:113:80c0:8080::4 dev br0  FAILED
    fe80::5054:ff:fef4:ecb8 dev br0 lladdr 52:54:00:f4:ec:b8 STALE
    fe80::208:2ff:feed:8f15 dev br0 lladdr 00:08:02:ed:8f:15 STALE
    fe80::2de:fbff:fee3:d77c dev br0 lladdr 00:de:fb:e3:d7:7c router STALE
    fe80::5054:ff:fe44:d768 dev br0 lladdr 52:54:00:44:d7:68 STALE
    fe80::501:abb4:eb5c:6686 dev br0 lladdr e4:b9:7a:e4:aa:ad STALE
    fe80::5054:ff:fe44:d767 dev br0 lladdr 52:54:00:44:d7:67 STALE
    fe80::5054:ff:fe87:8cc4 dev br0 lladdr 52:54:00:87:8c:c4 STALE
    fe80::c3ab:62d0:2723:6249 dev br0 lladdr 64:00:6a:75:00:72 STALE
    fe80::5054:ff:fe44:d766 dev br0 lladdr 52:54:00:44:d7:66 STALE
    fe80::800:ff:fe00:32 dev br0 lladdr 0a:00:00:00:00:32 STALE
    fe80::1a66:daff:fe00:bbaa dev br0 lladdr 18:66:da:00:bb:aa STALE
    fe80::36ac:19a7:3193:7081 dev br0 lladdr 0a:00:00:00:00:33 STALE
    fe80::216:3eff:fe48:17ff dev br0 lladdr 00:16:3e:48:17:ff STALE
    fe80::5054:ff:fe44:d765 dev br0 lladdr 52:54:00:44:d7:65 STALE
openqaworker5.suse.de:
    fe80::5054:ff:fe30:a4d9 dev eth0 lladdr 52:54:00:30:a4:d9 STALE
    fe80::6600:6aff:fe75:72 dev eth0 lladdr 64:00:6a:75:00:72 STALE
    fe80::208:2ff:feed:8f15 dev eth0 lladdr 00:08:02:ed:8f:15 STALE
    fe80::4950:d671:f08c:c9c3 dev eth0 lladdr 18:db:f2:46:1e:1d STALE
    fe80::5054:ff:fe44:d765 dev eth0 lladdr 52:54:00:44:d7:65 STALE
    fe80::800:ff:fe00:32 dev eth0 lladdr 0a:00:00:00:00:32 STALE
    fe80::9249:faff:fe06:82d8 dev eth0 lladdr 90:49:fa:06:82:d8 STALE
    fe80::d681:d7ff:fe5a:a39c dev eth0 lladdr d4:81:d7:5a:a3:9c STALE
    fe80::216:3eff:fe48:17ff dev eth0 lladdr 00:16:3e:48:17:ff STALE
    fe80::1a66:daff:fe00:bbaa dev eth0 lladdr 18:66:da:00:bb:aa STALE
    fe80::5054:ff:fe44:d767 dev eth0 lladdr 52:54:00:44:d7:67 STALE
    fe80::a3c9:d83f:17aa:8999 dev eth0 lladdr d4:81:d7:5a:a3:9c STALE
    fe80::5054:ff:fe29:137f dev eth0 lladdr 52:54:00:29:13:7f STALE
    fe80::36ac:19a7:3193:7081 dev eth0 lladdr 0a:00:00:00:00:33 STALE
    fe80::2af1:eff:fe41:cef3 dev eth0 lladdr 28:f1:0e:41:ce:f3 STALE
    fe80::5054:ff:fe87:8cc4 dev eth0 lladdr 52:54:00:87:8c:c4 STALE
    fe80::56ab:3aff:fe16:ddc4 dev eth0 lladdr 54:ab:3a:16:dd:c4 router STALE
    fe80::501:abb4:eb5c:6686 dev eth0 lladdr e4:b9:7a:e4:aa:ad STALE
    fe80::5054:ff:fe44:d766 dev eth0 lladdr 52:54:00:44:d7:66 STALE
    fe80::5054:ff:feb1:4de dev eth0 lladdr 52:54:00:b1:04:de STALE
    fe80::5054:ff:fef4:ecb8 dev eth0 lladdr 52:54:00:f4:ec:b8 STALE
    fe80::1 dev eth0 lladdr 00:00:5e:00:02:02 router STALE
    2620:113:80c0:8080::4 dev eth0  FAILED
    fe80::800:ff:fe00:15 dev eth0 lladdr 0a:00:00:00:00:15 STALE
    fe80::2de:fbff:fee3:dafc dev eth0 lladdr 00:de:fb:e3:da:fc router STALE
    fe80::c3ab:62d0:2723:6249 dev eth0 lladdr 64:00:6a:75:00:72 STALE
    fe80::5054:ff:fe44:d768 dev eth0 lladdr 52:54:00:44:d7:68 STALE
    fe80::2de:fbff:fee3:d77c dev eth0 lladdr 00:de:fb:e3:d7:7c router STALE
    fe80::ec4:7aff:fe7a:7736 dev eth0 lladdr 0c:c4:7a:7a:77:36 STALE
    fe80::2908:884f:5368:dda dev eth0 lladdr c8:f7:50:40:f4:69 STALE
grenache-1.qa.suse.de:
openqaworker10.suse.de:
    fe80::c3ab:62d0:2723:6249 dev eth0 lladdr 64:00:6a:75:00:72 STALE
    fe80::800:ff:fe00:15 dev eth0 lladdr 0a:00:00:00:00:15 STALE
    fe80::5054:ff:fe44:d768 dev eth0 lladdr 52:54:00:44:d7:68 STALE
    fe80::4950:d671:f08c:c9c3 dev eth0 lladdr 18:db:f2:46:1e:1d STALE
    fe80::501:abb4:eb5c:6686 dev eth0 lladdr e4:b9:7a:e4:aa:ad STALE
    fe80::800:ff:fe00:32 dev eth0 lladdr 0a:00:00:00:00:32 STALE
    fe80::2de:fbff:fee3:dafc dev eth0 lladdr 00:de:fb:e3:da:fc router STALE
    fe80::208:2ff:feed:8f15 dev eth0 lladdr 00:08:02:ed:8f:15 STALE
    2620:113:80c0:8080::5 dev eth0  FAILED
    fe80::9249:faff:fe06:82d8 dev eth0 lladdr 90:49:fa:06:82:d8 STALE
    fe80::1 dev eth0 lladdr 00:00:5e:00:02:02 router STALE
    fe80::5054:ff:fe44:d765 dev eth0 lladdr 52:54:00:44:d7:65 STALE
    fe80::1a66:daff:fe00:bbaa dev eth0 lladdr 18:66:da:00:bb:aa STALE
    fe80::6600:6aff:fe75:72 dev eth0 lladdr 64:00:6a:75:00:72 STALE
    fe80::5054:ff:fe44:d767 dev eth0 lladdr 52:54:00:44:d7:67 STALE
    fe80::5054:ff:fef4:ecb8 dev eth0 lladdr 52:54:00:f4:ec:b8 STALE
    fe80::5054:ff:fe29:137f dev eth0 lladdr 52:54:00:29:13:7f STALE
    fe80::20d:b9ff:fe01:ea8 dev gre_sys  FAILED
    fe80::2de:fbff:fee3:d77c dev eth0 lladdr 00:de:fb:e3:d7:7c router STALE
    fe80::a3c9:d83f:17aa:8999 dev eth0 lladdr d4:81:d7:5a:a3:9c STALE
    fe80::5054:ff:fe30:a4d9 dev eth0 lladdr 52:54:00:30:a4:d9 STALE
    fe80::d681:d7ff:fe5a:a39c dev eth0 lladdr d4:81:d7:5a:a3:9c STALE
    fe80::216:3eff:fe48:17ff dev eth0 lladdr 00:16:3e:48:17:ff STALE
    fe80::5054:ff:fe87:8cc4 dev eth0 lladdr 52:54:00:87:8c:c4 STALE
    fe80::56ab:3aff:fe16:ddc4 dev eth0 lladdr 54:ab:3a:16:dd:c4 router STALE
    fe80::2af1:eff:fe41:cef3 dev eth0 lladdr 28:f1:0e:41:ce:f3 STALE
    fe80::36ac:19a7:3193:7081 dev eth0 lladdr 0a:00:00:00:00:33 STALE
    fe80::5054:ff:fe44:d766 dev eth0 lladdr 52:54:00:44:d7:66 STALE
    fe80::ec4:7aff:fe7a:7736 dev eth0 lladdr 0c:c4:7a:7a:77:36 STALE
openqaworker13.suse.de:
    fe80::5054:ff:fe44:d766 dev eth0 lladdr 52:54:00:44:d7:66 STALE
    fe80::1 dev eth0 lladdr 00:00:5e:00:02:02 router STALE
    fe80::800:ff:fe00:32 dev eth0 lladdr 0a:00:00:00:00:32 STALE
    fe80::6600:6aff:fe75:72 dev eth0 lladdr 64:00:6a:75:00:72 STALE
    fe80::2af1:eff:fe41:cef3 dev eth0 lladdr 28:f1:0e:41:ce:f3 STALE
    fe80::c3ab:62d0:2723:6249 dev eth0 lladdr 64:00:6a:75:00:72 STALE
    fe80::36ac:19a7:3193:7081 dev eth0 lladdr 0a:00:00:00:00:33 STALE
    fe80::ec4:7aff:fe7a:7736 dev eth0 lladdr 0c:c4:7a:7a:77:36 STALE
    fe80::5054:ff:fe44:d767 dev eth0 lladdr 52:54:00:44:d7:67 STALE
    fe80::216:3eff:fe48:17ff dev eth0 lladdr 00:16:3e:48:17:ff STALE
    2620:113:80c0:8080::5 dev eth0  FAILED
    fe80::9249:faff:fe06:82d8 dev eth0 lladdr 90:49:fa:06:82:d8 STALE
    fe80::5054:ff:fe44:d768 dev eth0 lladdr 52:54:00:44:d7:68 STALE
    fe80::5054:ff:fe30:a4d9 dev eth0 lladdr 52:54:00:30:a4:d9 STALE
    fe80::5054:ff:fef4:ecb8 dev eth0 lladdr 52:54:00:f4:ec:b8 STALE
    fe80::5054:ff:fe29:137f dev eth0 lladdr 52:54:00:29:13:7f STALE
    fe80::2de:fbff:fee3:dafc dev eth0 lladdr 00:de:fb:e3:da:fc router STALE
    fe80::4950:d671:f08c:c9c3 dev eth0 lladdr 18:db:f2:46:1e:1d STALE
    fe80::5054:ff:fe44:d765 dev eth0 lladdr 52:54:00:44:d7:65 STALE
    fe80::208:2ff:feed:8f15 dev eth0 lladdr 00:08:02:ed:8f:15 STALE
    fe80::800:ff:fe00:15 dev eth0 lladdr 0a:00:00:00:00:15 STALE
    fe80::56ab:3aff:fe16:ddc4 dev eth0 lladdr 54:ab:3a:16:dd:c4 router STALE
    fe80::501:abb4:eb5c:6686 dev eth0 lladdr e4:b9:7a:e4:aa:ad STALE
    fe80::5054:ff:fe87:8cc4 dev eth0 lladdr 52:54:00:87:8c:c4 STALE
    fe80::2de:fbff:fee3:d77c dev eth0 lladdr 00:de:fb:e3:d7:7c router STALE
    fe80::1a66:daff:fe00:bbaa dev eth0 lladdr 18:66:da:00:bb:aa STALE
openqaworker-arm-1.suse.de:
openqaworker-arm-2.suse.de:
QA-Power8-5-kvm.qa.suse.de:
    Minion did not return. [Not connected]
malbec.arch.suse.de:
    Minion did not return. [Not connected]
openqaworker-arm-3.suse.de:
    Minion did not return. [Not connected]

#10 Updated by okurz 12 months ago

nicksinger in https://infra.nue.suse.com/SelfService/Display.html?id=178626 mmaher asked the question "Did the operation with the s390 host in the qa network helped in this issue? is it still the case? or any other news?". Something is certainly still wrong but I think what we could do is to provide "steps to reproduce" in EngInfra tickets. Otherwise the poor lads and lassies really do not have a better chance then to ask the reporter "is it still happening". And here I am not even super sure. So is the way to test: "Reboot worker machine, make sure no workaround disables IPv6 and call ping6 -c 1 www.opensuse.org to check if IPv6 works?" or is ping6 -c 1 openqa.suse.de enough?

#11 Updated by okurz 12 months ago

  • Related to action #76828: big job queue for ppc as powerqaworker-qam-1.qa and malbec.arch and qa-power8-5-kvm were not active added

#12 Updated by nicksinger 12 months ago

okurz wrote:

nicksinger in https://infra.nue.suse.com/SelfService/Display.html?id=178626 mmaher asked the question "Did the operation with the s390 host in the qa network helped in this issue? is it still the case? or any other news?". Something is certainly still wrong but I think what we could do is to provide "steps to reproduce" in EngInfra tickets. Otherwise the poor lads and lassies really do not have a better chance then to ask the reporter "is it still happening". And here I am not even super sure. So is the way to test: "Reboot worker machine, make sure no workaround disables IPv6 and call ping6 -c 1 www.opensuse.org to check if IPv6 works?" or is ping6 -c 1 openqa.suse.de enough?

strictly speaking about the v6 issue I think your first approach is the best. It should also be possible to do all this "at runtime" but safest is a reboot of course.

#13 Updated by nicksinger 12 months ago

the repair of powerqaworker-qam-1 showed some interesting results as the machine was broken long enough to not get the most recent salt updates. Right after the machine was started:

powerqaworker-qam-1:~ # ip -6 r s
2620:113:80c0:80a0::/64 dev eth4 proto kernel metric 256 expires 3535sec pref medium
fe80::/64 dev br1 proto kernel metric 256 pref medium
fe80::/64 dev eth4 proto kernel metric 256 pref medium
default via fe80::1 dev eth4 proto ra metric 1024 expires 1735sec hoplimit 64 pref medium

at this time, the salt-key was blocklisted and therefore no states where applied. To conclude my work on https://progress.opensuse.org/issues/68053 I accepted the salt-key on OSD once again and issues an manual "state.highstate". Here is what was changed:

openqa:~ # salt 'powerqaworker-qam-1' state.highstate
powerqaworker-qam-1:
----------
          ID: firewalld
    Function: service.running
      Result: True
     Comment: Service firewalld is already enabled, and is running
     Started: 14:57:32.501786
    Duration: 545.143 ms
     Changes:
              ----------
              firewalld:
                  True
----------
          ID: grub-conf
    Function: augeas.change
      Result: True
     Comment: Changes have been saved
     Started: 14:57:37.141460
    Duration: 176.998 ms
     Changes:
              ----------
              diff:
                  ---
                  +++
                  @@ -14 +14 @@
                  -GRUB_CMDLINE_LINUX_DEFAULT="nospec kvm.nested=1 kvm_intel.nested=1 kvm_amd.nested=1 kvm-arm.nested=1 crashkernel=210M"
                  +GRUB_CMDLINE_LINUX_DEFAULT=" nospec kvm.nested=1 kvm_intel.nested=1 kvm_amd.nested=1 kvm-arm.nested=1 crashkernel=210M"
----------
          ID: grub2-mkconfig > /boot/grub2/grub.cfg
    Function: cmd.run
      Result: True
     Comment: Command "grub2-mkconfig > /boot/grub2/grub.cfg" run
     Started: 14:57:37.321017
    Duration: 708.689 ms
     Changes:
              ----------
              pid:
                  30665
              retcode:
                  0
              stderr:
                  Generating grub configuration file ...
                  Found linux image: /boot/vmlinux-4.12.14-lp151.28.75-default
                  Found initrd image: /boot/initrd-4.12.14-lp151.28.75-default
                  Found linux image: /boot/vmlinux-4.12.14-lp151.28.48-default
                  Found initrd image: /boot/initrd-4.12.14-lp151.28.48-default
                  done
              stdout:
----------
          ID: telegraf
    Function: service.running
      Result: True
     Comment: Started Service telegraf
     Started: 14:57:38.276106
    Duration: 171.584 ms
     Changes:
              ----------
              telegraf:
                  True

Summary for powerqaworker-qam-1
--------------
Succeeded: 270 (changed=4)
Failed:      0
--------------
Total states run:     270
Total run time:    35.355 s

and afterwards:

powerqaworker-qam-1:~ # ip -6 r s
2620:113:80c0:80a0::/64 dev eth4 proto kernel metric 256 expires 3355sec pref medium
fe80::/64 dev br1 proto kernel metric 256 pref medium
fe80::/64 dev eth4 proto kernel metric 256 pref medium

so everything points to firewalld ATM. Disabling firewalld didn't bring the default route back. Will see if I can somehow restore a "working system" again to bisect where our firewalld behaves wrong.

#14 Updated by nicksinger 12 months ago

firewalld is certainly to blame here. I've collected the dump of ip6tables but that's too much for me to digest for today
EDIT: colorized diff of these two files can be found at https://w3.suse.de/~nsinger/diff.html

#15 Updated by okurz 12 months ago

  • Status changed from Feedback to In Progress
  • Assignee changed from okurz to nicksinger

Great news. Please continue the firewalld investigation.

#16 Updated by nicksinger 12 months ago

seems like firewalld was just the trigger. Currently following the hint that if net.ipv6.conf.all.forwarding = 1 is set then net.ipv6.conf.eth1.accept_ra needs to be set to 2 to accept RA's which seem to be the base for wickedd-dhcp6

#17 Updated by nicksinger 12 months ago

alright so my suspicion was confirmed. Something caused net.ipv6.conf.all.forwarding to be set to 1 - I assume this was implicitly done by firewalld. According to https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt :

accept_ra - INTEGER
    Accept Router Advertisements; autoconfigure using them.

    It also determines whether or not to transmit Router
    Solicitations. If and only if the functional setting is to
    accept Router Advertisements, Router Solicitations will be
    transmitted.

    Possible values are:
        0 Do not accept Router Advertisements.
        1 Accept Router Advertisements if forwarding is disabled.
        2 Overrule forwarding behaviour. Accept Router Advertisements
          even if forwarding is enabled.

    Functional default: enabled if local forwarding is disabled.
                disabled if local forwarding is enabled.

Therefore our workers didn't receive any RA from the NEXUS anymore resulting in dhcpv6 (from wicked) not being able to configure IPv6 properly any longer. That's why we saw proper configured link-local addresses (fe80::/64) but no link-global (2620:113:80c0:8080::/64 - this is the suse prefix). Also the default route over fe80::1 was missing because of the missing (or rather, not accepted) RA's.

BTW: I was able to reproduce the severe performance impact that we saw once we added fe80::1 manually as default route. This is caused if you only have a default route but no route for your prefix resulting in ICMP redirects from the router each and every time the machine tries to reach something in its own v6-subnet (which is basically every machine inside SUSE). This redirect resulted in a massive amount of re-transmitted TCP packages dropping the performance down to max 5MB/s and even stall connections for almost the rest of the time.

#18 Updated by nicksinger 12 months ago

This was the current (broken) state. Please note that worker8, qam-1, worker2 and both arms where my test subjects so it's expected to look correct there. All others show no default route for v6:

openqa:~ # salt -l error -C 'G@roles:worker' cmd.run 'ip -6 r s | grep default'
openqaworker3.suse.de:
openqaworker9.suse.de:
openqaworker8.suse.de:
    default via fe80::1 dev eth1 proto ra metric 1024 expires 3418sec hoplimit 64 pref medium
openqaworker6.suse.de:
QA-Power8-4-kvm.qa.suse.de:
powerqaworker-qam-1:
    default via fe80::1 dev eth4 metric 1024 pref medium
openqaworker5.suse.de:
QA-Power8-5-kvm.qa.suse.de:
openqaworker2.suse.de:
    default via fe80::1 dev br0 proto ra metric 1024 expires 3418sec hoplimit 64 pref medium
malbec.arch.suse.de:
grenache-1.qa.suse.de:
openqaworker13.suse.de:
openqaworker10.suse.de:
openqaworker-arm-1.suse.de:
    default via fe80::1 dev eth0 proto ra metric 1024 expires 3417sec hoplimit 64 pref medium
openqaworker-arm-2.suse.de:
    default via fe80::1 dev eth1 proto ra metric 1024 expires 3417sec hoplimit 64 pref medium

So what I did now to fix this is the following:

  1. net.ipv6.conf.all.disable_ipv6=0 to enable ipv6 on all interfaces again removing any previous workaround on the machines
  2. With $(ip r s | grep default | sed -n "s/^.*dev \(.*\) proto dhcp/\1/p" | xargs) I get the default interface for v4 traffic. Since we use the same interface for both address types we can just use it as default for all v6 operations that follow now
  3. sysctl net.ipv6.conf.$default_interface.disable_ipv6=1 disable v6 explicitly on the uplink so we can see if it worked afterwards
  4. sysctl net.ipv6.conf.$default_interface.accept_ra=2 enable RA's on the uplink only. We could set it for all interfaces but SUTs could misbehave and shouldn't affect the workers interface…
  5. sysctl net.ipv6.conf.$default_interface.disable_ipv6=0 bring back v6 to instantly trigger a SLAAC

The actual salt command looks a little messy but basically the steps described above:

openqa:~ # salt -l error -C 'G@roles:worker' cmd.run 'sysctl net.ipv6.conf.all.disable_ipv6=0; sysctl net.ipv6.conf.$(ip r s | grep default | sed -n "s/^.*dev \(.*\) proto dhcp/\1/p" | xargs).disable_ipv6=1; sysctl net.ipv6.conf.$(ip r s | grep default | sed -n "s/^.*dev \(.*\) proto dhcp/\1/p" | xargs).accept_ra=2; sysctl net.ipv6.conf.$(ip r s | grep default | sed -n "s/^.*dev \(.*\) proto dhcp/\1/p" | xargs).disable_ipv6=0'
openqaworker8.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth1.disable_ipv6 = 1
    net.ipv6.conf.eth1.accept_ra = 2
    net.ipv6.conf.eth1.disable_ipv6 = 0
openqaworker3.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.br0.disable_ipv6 = 1
    net.ipv6.conf.br0.accept_ra = 2
    net.ipv6.conf.br0.disable_ipv6 = 0
powerqaworker-qam-1:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth4.disable_ipv6 = 1
    net.ipv6.conf.eth4.accept_ra = 2
    net.ipv6.conf.eth4.disable_ipv6 = 0
QA-Power8-5-kvm.qa.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth3.disable_ipv6 = 1
    net.ipv6.conf.eth3.accept_ra = 2
    net.ipv6.conf.eth3.disable_ipv6 = 0
QA-Power8-4-kvm.qa.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth3.disable_ipv6 = 1
    net.ipv6.conf.eth3.accept_ra = 2
    net.ipv6.conf.eth3.disable_ipv6 = 0
malbec.arch.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth4.disable_ipv6 = 1
    net.ipv6.conf.eth4.accept_ra = 2
    net.ipv6.conf.eth4.disable_ipv6 = 0
grenache-1.qa.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth0.disable_ipv6 = 1
    net.ipv6.conf.eth0.accept_ra = 2
    net.ipv6.conf.eth0.disable_ipv6 = 0
openqaworker6.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth0.disable_ipv6 = 1
    net.ipv6.conf.eth0.accept_ra = 2
    net.ipv6.conf.eth0.disable_ipv6 = 0
openqaworker9.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth1.disable_ipv6 = 1
    net.ipv6.conf.eth1.accept_ra = 2
    net.ipv6.conf.eth1.disable_ipv6 = 0
openqaworker-arm-1.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth0.disable_ipv6 = 1
    net.ipv6.conf.eth0.accept_ra = 2
    net.ipv6.conf.eth0.disable_ipv6 = 0
openqaworker13.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth0.disable_ipv6 = 1
    net.ipv6.conf.eth0.accept_ra = 2
    net.ipv6.conf.eth0.disable_ipv6 = 0
openqaworker5.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth0.disable_ipv6 = 1
    net.ipv6.conf.eth0.accept_ra = 2
    net.ipv6.conf.eth0.disable_ipv6 = 0
openqaworker-arm-2.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth1.disable_ipv6 = 1
    net.ipv6.conf.eth1.accept_ra = 2
    net.ipv6.conf.eth1.disable_ipv6 = 0
openqaworker2.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.br0.disable_ipv6 = 1
    net.ipv6.conf.br0.accept_ra = 2
    net.ipv6.conf.br0.disable_ipv6 = 0
openqaworker10.suse.de:
    net.ipv6.conf.all.disable_ipv6 = 0
    net.ipv6.conf.eth0.disable_ipv6 = 1
    net.ipv6.conf.eth0.accept_ra = 2
    net.ipv6.conf.eth0.disable_ipv6 = 0

After I issued the command from above:

openqa:~ # salt -l error -C 'G@roles:worker' cmd.run 'ip -6 r s | grep default'
openqaworker3.suse.de:
    default via fe80::1 dev br0 proto ra metric 1024 expires 3491sec hoplimit 64 pref medium
openqaworker8.suse.de:
    default via fe80::1 dev eth1 proto ra metric 1024 expires 3493sec hoplimit 64 pref medium
openqaworker5.suse.de:
    default via fe80::1 dev eth0 proto ra metric 1024 expires 3493sec hoplimit 64 pref medium
openqaworker9.suse.de:
    default via fe80::1 dev eth1 proto ra metric 1024 expires 3493sec hoplimit 64 pref medium
openqaworker2.suse.de:
    default via fe80::1 dev br0 proto ra metric 1024 expires 3494sec hoplimit 64 pref medium
QA-Power8-5-kvm.qa.suse.de:
    default via fe80::1 dev eth3 proto ra metric 1024 expires 1691sec hoplimit 64 pref medium
openqaworker6.suse.de:
    default via fe80::1 dev eth0 proto ra metric 1024 expires 3493sec hoplimit 64 pref medium
powerqaworker-qam-1:
    default via fe80::1 dev eth4 proto ra metric 1024 expires 1692sec hoplimit 64 pref medium
QA-Power8-4-kvm.qa.suse.de:
    default via fe80::1 dev eth3 proto ra metric 1024 expires 1690sec hoplimit 64 pref medium
malbec.arch.suse.de:
    default via fe80::1 dev eth4 proto ra metric 1024 expires 3502sec hoplimit 64 pref medium
grenache-1.qa.suse.de:
    default via fe80::1 dev eth0 proto ra metric 1024 expires 1691sec hoplimit 64 pref medium
openqaworker10.suse.de:
    default via fe80::1 dev eth0 proto ra metric 1024 expires 3493sec hoplimit 64 pref medium
openqaworker13.suse.de:
    default via fe80::1 dev eth0 proto ra metric 1024 expires 3493sec hoplimit 64 pref medium
openqaworker-arm-1.suse.de:
    default via fe80::1 dev eth0 proto ra metric 1024 expires 3492sec hoplimit 64 pref medium
openqaworker-arm-2.suse.de:
    default via fe80::1 dev eth1 proto ra metric 1024 expires 3493sec hoplimit 64 pref medium

#19 Updated by nicksinger 12 months ago

After applying these changes, OSD can be reached over v6 from all machines:

openqa:~ # salt -l error -C 'G@roles:worker' cmd.run 'ping6 -c 1 openqa.suse.de'
openqaworker2.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=64 time=0.281 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.281/0.281/0.281/0.000 ms
openqaworker8.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=64 time=0.664 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.664/0.664/0.664/0.000 ms
openqaworker3.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=64 time=0.496 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.496/0.496/0.496/0.000 ms
openqaworker6.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=64 time=0.167 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.167/0.167/0.167/0.000 ms
openqaworker9.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=64 time=0.381 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.381/0.381/0.381/0.000 ms
QA-Power8-5-kvm.qa.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=63 time=0.278 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.278/0.278/0.278/0.000 ms
openqaworker5.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=64 time=0.614 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.614/0.614/0.614/0.000 ms
powerqaworker-qam-1:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=63 time=0.214 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.214/0.214/0.214/0.000 ms
QA-Power8-4-kvm.qa.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=63 time=0.197 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.197/0.197/0.197/0.000 ms
malbec.arch.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=63 time=0.183 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.183/0.183/0.183/0.000 ms
grenache-1.qa.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=63 time=0.478 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.478/0.478/0.478/0.000 ms
openqaworker10.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=64 time=0.154 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.154/0.154/0.154/0.000 ms
openqaworker13.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=64 time=0.236 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.236/0.236/0.236/0.000 ms
openqaworker-arm-1.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=64 time=0.297 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 0.297/0.297/0.297/0.000 ms
openqaworker-arm-2.suse.de:
    PING openqa.suse.de(openqa.suse.de (2620:113:80c0:8080:10:160:0:207)) 56 data bytes
    64 bytes from openqa.suse.de (2620:113:80c0:8080:10:160:0:207): icmp_seq=1 ttl=64 time=3.09 ms

    --- openqa.suse.de ping statistics ---
    1 packets transmitted, 1 received, 0% packet loss, time 0ms
    rtt min/avg/max/mdev = 3.090/3.090/3.090/0.000 ms

Just because we saw performance issues with the last workaround I deployed I wanted to have a speedtest too. The other side was running on my workstation which is in the same VLAN but is at least always a hop (office switches) away from the workers:

openqa:~ # salt -b 1 -l error -C 'G@roles:worker' cmd.run 'which iperf3 && iperf3 -c 2620:113:80c0:80a0:10:162:32:1f7'

Executing run on ['openqaworker2.suse.de']

jid:
    20201106124828227894
openqaworker2.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:8080:2e60:cff:fe73:2ac port 51558 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   102 MBytes   855 Mbits/sec  444   32.1 KBytes
    [  5]   1.00-2.00   sec  97.0 MBytes   814 Mbits/sec  338   18.1 KBytes
    [  5]   2.00-3.00   sec  95.6 MBytes   802 Mbits/sec  721    312 KBytes
    [  5]   3.00-4.00   sec  96.4 MBytes   809 Mbits/sec  628   34.9 KBytes
    [  5]   4.00-5.00   sec  92.2 MBytes   773 Mbits/sec  301   34.9 KBytes
    [  5]   5.00-6.00   sec  89.3 MBytes   749 Mbits/sec  494    113 KBytes
    [  5]   6.00-7.00   sec  87.3 MBytes   733 Mbits/sec  609    106 KBytes
    [  5]   7.00-8.00   sec  87.0 MBytes   730 Mbits/sec  325    251 KBytes
    [  5]   8.00-9.00   sec  86.3 MBytes   724 Mbits/sec  246   83.7 KBytes
    [  5]   9.00-10.00  sec  73.2 MBytes   614 Mbits/sec   93    142 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec   906 MBytes   760 Mbits/sec  4199             sender
    [  5]   0.00-10.04  sec   905 MBytes   756 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

Executing run on ['powerqaworker-qam-1']

jid:
    20201106124838673989
powerqaworker-qam-1:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:80a0:10:162:30:de72 port 60628 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   104 MBytes   876 Mbits/sec   19    279 KBytes
    [  5]   1.00-2.00   sec   105 MBytes   881 Mbits/sec    2    286 KBytes
    [  5]   2.00-3.00   sec   105 MBytes   881 Mbits/sec   14    258 KBytes
    [  5]   3.00-4.00   sec   105 MBytes   881 Mbits/sec    6    255 KBytes
    [  5]   4.00-5.00   sec   106 MBytes   891 Mbits/sec    5    297 KBytes
    [  5]   5.00-6.00   sec   100 MBytes   839 Mbits/sec    8    252 KBytes
    [  5]   6.00-7.00   sec   108 MBytes   902 Mbits/sec    3    280 KBytes
    [  5]   7.00-8.00   sec   102 MBytes   860 Mbits/sec    5    322 KBytes
    [  5]   8.00-9.00   sec   108 MBytes   902 Mbits/sec    7    303 KBytes
    [  5]   9.00-10.00  sec   109 MBytes   912 Mbits/sec   10    261 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec  1.03 GBytes   882 Mbits/sec   79             sender
    [  5]   0.00-10.04  sec  1.02 GBytes   876 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

Executing run on ['openqaworker13.suse.de']

jid:
    20201106124849020759
openqaworker13.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:8080:10:160:2:26 port 53016 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   106 MBytes   893 Mbits/sec   27    247 KBytes
    [  5]   1.00-2.00   sec   107 MBytes   897 Mbits/sec   73    218 KBytes
    [  5]   2.00-3.00   sec   108 MBytes   902 Mbits/sec   83    204 KBytes
    [  5]   3.00-4.00   sec   103 MBytes   867 Mbits/sec   34    251 KBytes
    [  5]   4.00-5.00   sec   103 MBytes   866 Mbits/sec   55    132 KBytes
    [  5]   5.00-6.00   sec   105 MBytes   880 Mbits/sec   63    230 KBytes
    [  5]   6.00-7.00   sec   101 MBytes   849 Mbits/sec  129    201 KBytes
    [  5]   7.00-8.00   sec   104 MBytes   869 Mbits/sec   26    114 KBytes
    [  5]   8.00-9.00   sec   104 MBytes   873 Mbits/sec   60    218 KBytes
    [  5]   9.00-10.00  sec   103 MBytes   865 Mbits/sec   51    211 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec  1.02 GBytes   876 Mbits/sec  601             sender
    [  5]   0.00-10.04  sec  1.02 GBytes   870 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

Executing run on ['openqaworker10.suse.de']

jid:
    20201106124859469998
openqaworker10.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:8080:10:160:68:1 port 38128 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   111 MBytes   930 Mbits/sec   51    208 KBytes
    [  5]   1.00-2.00   sec   107 MBytes   900 Mbits/sec  132    199 KBytes
    [  5]   2.00-3.00   sec   106 MBytes   892 Mbits/sec  126   69.7 KBytes
    [  5]   3.00-4.00   sec   106 MBytes   891 Mbits/sec  115    159 KBytes
    [  5]   4.00-5.00   sec   105 MBytes   879 Mbits/sec  125    279 KBytes
    [  5]   5.00-6.00   sec   109 MBytes   911 Mbits/sec   75    252 KBytes
    [  5]   6.00-7.00   sec   104 MBytes   869 Mbits/sec  124    291 KBytes
    [  5]   7.00-8.00   sec   107 MBytes   894 Mbits/sec  130    216 KBytes
    [  5]   8.00-9.00   sec   109 MBytes   915 Mbits/sec   42    199 KBytes
    [  5]   9.00-10.00  sec   107 MBytes   898 Mbits/sec  128    223 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec  1.05 GBytes   898 Mbits/sec  1048             sender
    [  5]   0.00-10.05  sec  1.04 GBytes   892 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

Executing run on ['QA-Power8-5-kvm.qa.suse.de']

QA-Power8-5-kvm.qa.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:80a0:10:162:2a:5c8d port 52542 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec  89.5 MBytes   751 Mbits/sec  327    325 KBytes
    [  5]   1.00-2.00   sec  92.0 MBytes   772 Mbits/sec  624    250 KBytes
    [  5]   2.00-3.00   sec  98.5 MBytes   826 Mbits/sec  490   73.9 KBytes
    [  5]   3.00-4.00   sec  94.6 MBytes   793 Mbits/sec  607    152 KBytes
    [  5]   4.00-5.00   sec  96.2 MBytes   807 Mbits/sec  521    445 KBytes
    [  5]   5.00-6.00   sec  95.7 MBytes   803 Mbits/sec  833   34.9 KBytes
    [  5]   6.00-7.01   sec  95.8 MBytes   799 Mbits/sec  787   78.1 KBytes
    [  5]   7.01-8.00   sec  89.4 MBytes   755 Mbits/sec  980    181 KBytes
    [  5]   8.00-9.00   sec  91.4 MBytes   767 Mbits/sec  243    137 KBytes
    [  5]   9.00-10.00  sec  73.5 MBytes   616 Mbits/sec  515   25.1 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec   917 MBytes   769 Mbits/sec  5927             sender
    [  5]   0.00-10.04  sec   914 MBytes   764 Mbits/sec                  receiver

    iperf Done.
jid:
    20201106124909926319
retcode:
    0

Executing run on ['openqaworker5.suse.de']

jid:
    20201106124920344427
openqaworker5.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:8080:10:160:1:93 port 50440 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   105 MBytes   877 Mbits/sec  309    107 KBytes
    [  5]   1.00-2.00   sec  99.8 MBytes   837 Mbits/sec  337    110 KBytes
    [  5]   2.00-3.00   sec   103 MBytes   868 Mbits/sec  154    188 KBytes
    [  5]   3.00-4.00   sec  99.8 MBytes   837 Mbits/sec  377    314 KBytes
    [  5]   4.00-5.00   sec   100 MBytes   843 Mbits/sec  432   86.5 KBytes
    [  5]   5.00-6.00   sec  99.5 MBytes   835 Mbits/sec  310    234 KBytes
    [  5]   6.00-7.00   sec   104 MBytes   872 Mbits/sec  222    206 KBytes
    [  5]   7.00-8.00   sec  99.5 MBytes   834 Mbits/sec  246    107 KBytes
    [  5]   8.00-9.00   sec  98.8 MBytes   829 Mbits/sec  290    251 KBytes
    [  5]   9.00-10.00  sec  96.6 MBytes   811 Mbits/sec  465    155 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec  1006 MBytes   844 Mbits/sec  3142             sender
    [  5]   0.00-10.04  sec  1004 MBytes   839 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

Executing run on ['openqaworker8.suse.de']

jid:
    20201106124930709117
openqaworker8.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:8080:ec4:7aff:fe99:dc5b port 54914 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec  96.2 MBytes   807 Mbits/sec  824   26.5 KBytes
    [  5]   1.00-2.00   sec  94.3 MBytes   791 Mbits/sec  404    160 KBytes
    [  5]   2.00-3.00   sec  87.8 MBytes   737 Mbits/sec  510   26.5 KBytes
    [  5]   3.00-4.00   sec  95.4 MBytes   800 Mbits/sec  709    230 KBytes
    [  5]   4.00-5.00   sec  98.5 MBytes   827 Mbits/sec  604    127 KBytes
    [  5]   5.00-6.00   sec  93.0 MBytes   780 Mbits/sec  709   32.1 KBytes
    [  5]   6.00-7.00   sec  97.8 MBytes   820 Mbits/sec  419   75.3 KBytes
    [  5]   7.00-8.00   sec  94.6 MBytes   793 Mbits/sec  605   93.4 KBytes
    [  5]   8.00-9.00   sec   102 MBytes   853 Mbits/sec  484    244 KBytes
    [  5]   9.00-10.00  sec  46.6 MBytes   391 Mbits/sec   78   60.0 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec   906 MBytes   760 Mbits/sec  5346             sender
    [  5]   0.00-10.04  sec   904 MBytes   755 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

Executing run on ['openqaworker9.suse.de']

jid:
    20201106124941047009
openqaworker9.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:8080:10:160:1:20 port 34090 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec  57.0 MBytes   478 Mbits/sec    3    713 KBytes
    [  5]   1.00-2.00   sec  55.0 MBytes   461 Mbits/sec   90    307 KBytes
    [  5]   2.00-3.00   sec   101 MBytes   849 Mbits/sec  109   71.1 KBytes
    [  5]   3.00-4.00   sec   100 MBytes   839 Mbits/sec  522    293 KBytes
    [  5]   4.00-5.00   sec   102 MBytes   860 Mbits/sec  212    211 KBytes
    [  5]   5.00-6.00   sec   101 MBytes   849 Mbits/sec  342    269 KBytes
    [  5]   6.00-7.00   sec  86.2 MBytes   724 Mbits/sec  499    276 KBytes
    [  5]   7.00-8.00   sec  48.8 MBytes   409 Mbits/sec  1401   37.7 KBytes
    [  5]   8.00-9.00   sec  71.2 MBytes   598 Mbits/sec  576    170 KBytes
    [  5]   9.00-10.00  sec  96.2 MBytes   807 Mbits/sec  640    218 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec   820 MBytes   687 Mbits/sec  4394             sender
    [  5]   0.00-10.04  sec   816 MBytes   682 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

Executing run on ['QA-Power8-4-kvm.qa.suse.de']

QA-Power8-4-kvm.qa.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:80a0:10:162:31:3446 port 53754 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   106 MBytes   889 Mbits/sec   17    180 KBytes
    [  5]   1.00-2.00   sec   101 MBytes   843 Mbits/sec    1    377 KBytes
    [  5]   2.00-3.00   sec   100 MBytes   842 Mbits/sec   35    198 KBytes
    [  5]   3.00-4.00   sec   104 MBytes   871 Mbits/sec   19   78.1 KBytes
    [  5]   4.00-5.00   sec   102 MBytes   859 Mbits/sec   18    322 KBytes
    [  5]   5.00-6.00   sec  84.8 MBytes   711 Mbits/sec    2    282 KBytes
    [  5]   6.00-7.00   sec  89.4 MBytes   750 Mbits/sec   14    257 KBytes
    [  5]   7.00-8.00   sec   103 MBytes   860 Mbits/sec    8    279 KBytes
    [  5]   8.00-9.00   sec   100 MBytes   843 Mbits/sec    9    298 KBytes
    [  5]   9.00-10.00  sec   104 MBytes   876 Mbits/sec    6    245 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec   995 MBytes   834 Mbits/sec  129             sender
    [  5]   0.00-10.06  sec   992 MBytes   828 Mbits/sec                  receiver

    iperf Done.
jid:
    20201106124951627941
retcode:
    0

Executing run on ['openqaworker3.suse.de']

jid:
    20201106125002001609
openqaworker3.suse.de:
    which: no iperf3 in (/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin)
retcode:
    1

Executing run on ['grenache-1.qa.suse.de']

grenache-1.qa.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:80a0:10:162:29:12f0 port 43282 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   113 MBytes   952 Mbits/sec   15    169 KBytes
    [  5]   1.00-2.00   sec   110 MBytes   923 Mbits/sec    3    176 KBytes
    [  5]   2.00-3.00   sec   110 MBytes   923 Mbits/sec   10    294 KBytes
    [  5]   3.00-4.00   sec   110 MBytes   923 Mbits/sec   11    170 KBytes
    [  5]   4.00-5.00   sec   111 MBytes   933 Mbits/sec   11    329 KBytes
    [  5]   5.00-6.00   sec   109 MBytes   912 Mbits/sec   12    280 KBytes
    [  5]   6.00-7.00   sec   111 MBytes   933 Mbits/sec   15    153 KBytes
    [  5]   7.00-8.00   sec   110 MBytes   923 Mbits/sec   11    315 KBytes
    [  5]   8.00-9.00   sec   110 MBytes   923 Mbits/sec   11    163 KBytes
    [  5]   9.00-10.00  sec   110 MBytes   923 Mbits/sec   13    149 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec  1.08 GBytes   927 Mbits/sec  112             sender
    [  5]   0.00-10.05  sec  1.08 GBytes   920 Mbits/sec                  receiver

    iperf Done.
jid:
    20201106125002233677
retcode:
    0

Executing run on ['openqaworker-arm-2.suse.de']

jid:
    20201106125012911475
openqaworker-arm-2.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:8080:1e1b:dff:fe68:ee4d port 48828 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   112 MBytes   936 Mbits/sec    0   3.00 MBytes
    [  5]   1.00-2.00   sec   109 MBytes   912 Mbits/sec    0   3.00 MBytes
    [  5]   2.00-3.00   sec   109 MBytes   912 Mbits/sec    0   3.00 MBytes
    [  5]   3.00-4.00   sec   109 MBytes   912 Mbits/sec    0   3.00 MBytes
    [  5]   4.00-5.00   sec   106 MBytes   892 Mbits/sec    0   3.00 MBytes
    [  5]   5.00-6.00   sec  98.8 MBytes   828 Mbits/sec    0   3.00 MBytes
    [  5]   6.00-7.00   sec   110 MBytes   923 Mbits/sec    0   3.00 MBytes
    [  5]   7.00-8.00   sec   110 MBytes   923 Mbits/sec    0   3.00 MBytes
    [  5]   8.00-9.00   sec   111 MBytes   933 Mbits/sec    0   3.15 MBytes
    [  5]   9.00-10.00  sec   110 MBytes   923 Mbits/sec    0   3.15 MBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec  1.06 GBytes   909 Mbits/sec    0             sender
    [  5]   0.00-10.03  sec  1.06 GBytes   907 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

Executing run on ['openqaworker-arm-1.suse.de']

jid:
    20201106125023654762
openqaworker-arm-1.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:8080:1e1b:dff:fe68:7ec7 port 60232 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   114 MBytes   954 Mbits/sec    0   3.00 MBytes
    [  5]   1.00-2.00   sec   106 MBytes   892 Mbits/sec    0   3.00 MBytes
    [  5]   2.00-3.00   sec   108 MBytes   902 Mbits/sec    0   3.00 MBytes
    [  5]   3.00-4.00   sec   108 MBytes   902 Mbits/sec    0   3.00 MBytes
    [  5]   4.00-5.00   sec   109 MBytes   912 Mbits/sec    0   3.00 MBytes
    [  5]   5.00-6.00   sec   110 MBytes   923 Mbits/sec    0   3.00 MBytes
    [  5]   6.00-7.00   sec   109 MBytes   912 Mbits/sec    0   3.00 MBytes
    [  5]   7.00-8.00   sec   109 MBytes   912 Mbits/sec    0   3.00 MBytes
    [  5]   8.00-9.00   sec   109 MBytes   912 Mbits/sec    0   3.00 MBytes
    [  5]   9.00-10.00  sec   110 MBytes   923 Mbits/sec    0   3.00 MBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec  1.06 GBytes   914 Mbits/sec    0             sender
    [  5]   0.00-10.03  sec  1.06 GBytes   912 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

Executing run on ['malbec.arch.suse.de']

jid:
    20201106125034314896
malbec.arch.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:8000:10:161:24:54 port 38826 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   107 MBytes   894 Mbits/sec    7    148 KBytes
    [  5]   1.00-2.00   sec   105 MBytes   881 Mbits/sec    2    381 KBytes
    [  5]   2.00-3.00   sec   102 MBytes   860 Mbits/sec    3    259 KBytes
    [  5]   3.00-4.00   sec   104 MBytes   870 Mbits/sec    4    291 KBytes
    [  5]   4.00-5.00   sec   102 MBytes   860 Mbits/sec    5    322 KBytes
    [  5]   5.00-6.00   sec   108 MBytes   902 Mbits/sec    7    305 KBytes
    [  5]   6.00-7.00   sec   105 MBytes   881 Mbits/sec    2    296 KBytes
    [  5]   7.00-8.00   sec   105 MBytes   881 Mbits/sec    7    276 KBytes
    [  5]   8.00-9.00   sec   102 MBytes   860 Mbits/sec    6    284 KBytes
    [  5]   9.00-10.00  sec   101 MBytes   849 Mbits/sec    3    317 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec  1.02 GBytes   874 Mbits/sec   46             sender
    [  5]   0.00-10.05  sec  1.01 GBytes   867 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

Executing run on ['openqaworker6.suse.de']

jid:
    20201106125044714148
openqaworker6.suse.de:
    /usr/bin/iperf3
    Connecting to host 2620:113:80c0:80a0:10:162:32:1f7, port 5201
    [  5] local 2620:113:80c0:8080:10:160:1:100 port 45096 connected to 2620:113:80c0:80a0:10:162:32:1f7 port 5201
    [ ID] Interval           Transfer     Bitrate         Retr  Cwnd
    [  5]   0.00-1.00   sec   101 MBytes   847 Mbits/sec  252    141 KBytes
    [  5]   1.00-2.00   sec   101 MBytes   843 Mbits/sec  275    209 KBytes
    [  5]   2.00-3.00   sec  99.3 MBytes   833 Mbits/sec  349   99.0 KBytes
    [  5]   3.00-4.00   sec  96.2 MBytes   807 Mbits/sec  314    243 KBytes
    [  5]   4.00-5.00   sec   100 MBytes   841 Mbits/sec  424    144 KBytes
    [  5]   5.00-6.00   sec  79.6 MBytes   668 Mbits/sec  284    250 KBytes
    [  5]   6.00-7.00   sec  98.8 MBytes   829 Mbits/sec  327   93.4 KBytes
    [  5]   7.00-8.00   sec   101 MBytes   848 Mbits/sec  336    145 KBytes
    [  5]   8.00-9.00   sec  97.7 MBytes   820 Mbits/sec  468    106 KBytes
    [  5]   9.00-10.00  sec  97.6 MBytes   818 Mbits/sec  345    144 KBytes
    - - - - - - - - - - - - - - - - - - - - - - - - -
    [ ID] Interval           Transfer     Bitrate         Retr
    [  5]   0.00-10.00  sec   972 MBytes   815 Mbits/sec  3374             sender
    [  5]   0.00-10.04  sec   970 MBytes   810 Mbits/sec                  receiver

    iperf Done.
retcode:
    0

So with these numbers I'm pretty certain that everything works as expected.

#20 Updated by cdywan 12 months ago

  • Status changed from In Progress to Resolved

nicksinger wrote:

After applying these changes, OSD can be reached over v6 from all machines:
[...]
So with these numbers I'm pretty certain that everything works as expected.

So the ticket is Resolved I take it?

#21 Updated by okurz 12 months ago

  • Status changed from Resolved to In Progress

As this ticket was about an issue causing lots of problems and confusions but was also caused by the team itself I would really keep it open and up for the assignee to decide when it is "Resolved". Definitely I think an issue specific retrospective should be conducted

Also
https://infra.nue.suse.com/SelfService/Display.html?id=178626
is still open

#22 Updated by cdywan 12 months ago

  • Due date changed from 2020-10-24 to 2020-11-13

Ack

#23 Updated by nicksinger 12 months ago

besides what was mentioned by Oli we also need a proper permanent solution in salt

#24 Updated by okurz 12 months ago

As you wrote we need to set net.ipv6.conf.$main_interface.accept_ra = 2

To get $main_interface https://tedops.github.io/how-to-find-default-active-ethernet-interface.html looks promising, e.g. call

salt \* network.default_route inet

I guess in salt state files we should do:

net.ipv6.conf.{{ salt['network.default_route']('inet')[0]['interface'] }}.accept_ra:
  sysctl.present:
    - value: 2

if this does not work then probably a custom grain function should be used, as in https://lemarchand.io/saltstack-and-internal-network-interfaces/

#25 Updated by nicksinger 12 months ago

  • Description updated (diff)

#26 Updated by nicksinger 12 months ago

  • Description updated (diff)

#27 Updated by nicksinger 11 months ago

  • Has duplicate action #77995: worker instances on grenache-1 seem to fail (sometimes?) to connect to web-uis added

#28 Updated by okurz 11 months ago

Trying the suggestion from #73633#note-24 on osd with a temporary change to /srv/salt/openqa/worker.sls and trying to apply with salt 'openqaworker10*' state.apply test=True I get:

openqaworker10.suse.de:
    Data failed to compile:
----------
    Rendering SLS 'base:openqa.worker' failed: Jinja error: 'anycast' does not appear to be an IPv4 or IPv6 network
Traceback (most recent call last):
  File "/usr/lib/python3.6/site-packages/salt/utils/templates.py", line 394, in render_jinja_tmpl
    output = template.render(**decoded_context)
  File "/usr/lib/python3.6/site-packages/jinja2/asyncsupport.py", line 76, in render
    return original_render(self, *args, **kwargs)
  File "/usr/lib/python3.6/site-packages/jinja2/environment.py", line 1008, in render
    return self.environment.handle_exception(exc_info, True)
  File "/usr/lib/python3.6/site-packages/jinja2/environment.py", line 780, in handle_exception
    reraise(exc_type, exc_value, tb)
  File "/usr/lib/python3.6/site-packages/jinja2/_compat.py", line 37, in reraise
    raise value.with_traceback(tb)
  File "<template>", line 367, in top-level template code
  File "/usr/lib/python3.6/site-packages/salt/modules/network.py", line 1690, in default_route
    _routes = routes()
  File "/usr/lib/python3.6/site-packages/salt/modules/network.py", line 1647, in routes
    routes_ = _ip_route_linux()
  File "/usr/lib/python3.6/site-packages/salt/modules/network.py", line 569, in _ip_route_linux
    address_mask = convert_cidr(comps[0])
  File "/usr/lib/python3.6/site-packages/salt/modules/network.py", line 1149, in convert_cidr
    cidr = calc_net(cidr)
  File "/usr/lib/python3.6/site-packages/salt/modules/network.py", line 1171, in calc_net
    return salt.utils.network.calc_net(ip_addr, netmask)
  File "/usr/lib/python3.6/site-packages/salt/utils/network.py", line 1053, in calc_net
    return six.text_type(ipaddress.ip_network(ipaddr, strict=False))
  File "/usr/lib64/python3.6/ipaddress.py", line 84, in ip_network
    address)
ValueError: 'anycast' does not appear to be an IPv4 or IPv6 network

; line 367

#29 Updated by okurz 11 months ago

  • Related to action #68095: Migrate osd workers from SuSEfirewall2 to firewalld added

#30 Updated by okurz 11 months ago

Trying a "5 Whys" analysis.

First mkittler worked on migrating SuSEfirewall2 to firewalld in #68095 . On 2020-10-19 13:20 CEST the according salt change was deployed to all workers

We were informed about a "general problem" by our monitoring and also by user reports about 2h later. Even before 2020-10-20 12:46 CEST nicksinger has manually added routes to workers as described in #75055 which then caused further issues. This looked good because as #75055 states: "the worker appeared on all webui's again" but the performance decreased heavily and lead to #73633 .

Maybe there

  • Why did we not see any problems directly after the salt state was applied?

    • it was not "completely broken" and took 24h to trigger the big alert, likely just after nsinger applied additional changes
    • -> suggestion: We should have monitoring for basic "IPv4 and IPv6 works to ping, tcp, http from all machines to all machines". Make sure to explicitly select both stacks
    • -> suggestion: A passive performance measurement regarding throughput on interfaces
  • Why did we not already have a ticket for the issue that mmoese reported on 2020-10-20?

    • At the time we did not see "baremetal-support.qa.suse.de" as that important for us and could not link it to an issue in the general osd infrastructure.
    • -> suggestion: whenever we apply changes to the infrastructure we should have a ticket
    • TODO lookup the according infra ticket and check when it was created
    • -> suggestion: Whenever creating any external ticket, e.g. EngInfra, create internal tracker ticket. Because there might be more internal notes
  • Why did we not see the connection the firewalld migration #68095 ?

    • Because no tests directly linked to the ticket or deployed salt changes failed
    • -> suggestion: Same as in OSD deployment we should look for failed grafana
    • -> suggestion: Collect all the information between "last good" and "first bad" and then also find the git diff in openqa/salt-states-openqa
  • Why did mkittler and me think that the firewalld change was not the issue?

    • We thought firewalld was "long gone" because mkittler already created the SR at 2020-10-15 (but only partially deployed for better testing)
    • We jumped to the conclusion that IPv6 changes within the network out of our control should have triggered that
    • -> suggestion: Apply proper "scientific method" with written down hypotheses, experiments and conclusions in tickets, follow https://progress.opensuse.org/projects/openqav3/wiki#Further-decision-steps-working-on-test-issues
    • -> suggestion: Keep salt states to describe what should not be there
    • -> suggestion: Try out older btrfs snapshots in systems for crosschecking and boot with disabled salt. In the kernel cmdline append systemd.mask=salt-minion.service
  • Why did it take so long?

    • Because EngInfra was too slow to tell us it's not their fault
    • nicksinger did not get an answer for "long enough" so he figured it's our own fault
    • We thought "good enough workarounds are in place" and worked on other tickets that helped to resolve the actual issue, e.g. #75055 , #75016
    • -> Conclusion: Actually we did good because the user base was not impacted that much anymore, we had workarounds in place, we were investigating other issues but always kept the relation to this ticket in mind which in the end helped to fix it
  • Why are we still not finished?

    • Because cdywan does not run dailies to check on urgent tickets still open
    • -> suggestion: team should conduct a work backlog check on a daily base
    • We were not sure if any other person should take the ticket from nsinger
    • -> suggestion: nsinger does not mind if someone else provides a suggestion or takes over the ticket

-> #78127

#31 Updated by okurz 11 months ago

  • Copied to action #78127: follow-up to #73633 - lessons learned and suggestions added

#32 Updated by cdywan 11 months ago

  • Due date changed from 2020-11-13 to 2020-11-17

#33 Updated by nicksinger 11 months ago

I'll take over from here and will try to implement a proper salt solution. This is my plan of action :)

#34 Updated by nicksinger 11 months ago

  • Status changed from In Progress to Feedback

https://gitlab.suse.de/openqa/salt-states-openqa/-/merge_requests/401 - anyone disagrees that we can close this here once this MR is merged? :)

#35 Updated by okurz 11 months ago

well, the deploy pipeline failed so I suggest to resolve this ticket as soon as you can check that this setting is actually applied on all affected machines :)

And are there still workarounds in place that we need to remove?

#36 Updated by cdywan 11 months ago

okurz wrote:

well, the deploy pipeline failed so I suggest to resolve this ticket as soon as you can check that this setting is actually applied on all affected machines :)

And are there still workarounds in place that we need to remove?

I re-ran the pipeline on master and deploy failed like this:

RROR: Minions returned with non-zero exit code
1411openqaworker-arm-1.suse.de:
1412Summary for openqaworker-arm-1.suse.de
1413--------------
1414Succeeded: 285
1415Failed:      0

#37 Updated by okurz 11 months ago

If in the gitlab CI pipeline job log you scroll further to the top you can find https://gitlab.suse.de/openqa/salt-states-openqa/-/jobs/289706#L822 which says:

openqa-monitor.qa.suse.de:
    Data failed to compile:
----------
    Rendering SLS 'base:openqa.monitoring.grafana' failed: while constructing a mapping
  in "<unicode string>", line 10, column 1
found conflicting ID '/var/lib/grafana/dashboards//worker-localhost.json'
  in "<unicode string>", line 193, column 1
openqaworker2.suse.de:

which we have a ticket about: #75445 which seems to be causing more problems now, hence raising prio there.

#38 Updated by okurz 11 months ago

looked into the topic together with nsinger:
Experimented on openqaworker10:

# dig openqa.suse.de AAAA
dig: parse of /etc/resolv.conf failed

looking into /etc/resolv.conf which was from 2018-10-19 with content:

search suse.de
nameserver fe80::20d:b9ff:fe01:ea8%eth2
nameserver 10.160.0.1

calling netconfig update -f replaced the file with a symlink. I remember that there was some system upgrade where one should have replaced manually maintained files with symlinks like this. Probably a good idea to do that on all our machines.

Now we can test again properly:

# dig openqa.suse.de AAAA

; <<>> DiG 9.16.6 <<>> openqa.suse.de AAAA
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 12530
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 4, ADDITIONAL: 7

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
; COOKIE: 91ba81fa158dab233fb0c3735fb788ad2f77d9cbe5e72c49 (good)
;; QUESTION SECTION:
;openqa.suse.de.            IN  AAAA

;; ANSWER SECTION:
openqa.suse.de.     300 IN  AAAA    2620:113:80c0:8080:10:160:0:207

;; AUTHORITY SECTION:
suse.de.        300 IN  NS  dns1.suse.de.
suse.de.        300 IN  NS  frao-p-infoblox-01.corp.suse.com.
suse.de.        300 IN  NS  dns2.suse.de.
suse.de.        300 IN  NS  frao-p-infoblox-02.corp.suse.com.

;; ADDITIONAL SECTION:
dns2.suse.de.       300 IN  AAAA    2620:113:80c0:8080:10:160:0:1
dns1.suse.de.       300 IN  AAAA    2620:113:80c0:8080:10:160:2:88
dns2.suse.de.       300 IN  A   10.160.0.1
dns1.suse.de.       300 IN  A   10.160.2.88
frao-p-infoblox-02.corp.suse.com. 14863 IN A    10.156.86.70
frao-p-infoblox-01.corp.suse.com. 14863 IN A    10.156.86.6

;; Query time: 0 msec
;; SERVER: 2620:113:80c0:8080:10:160:0:1#53(2620:113:80c0:8080:10:160:0:1)
;; WHEN: Fri Nov 20 10:13:17 CET 2020
;; MSG SIZE  rcvd: 336
# which iperf3 && iperf3 -c 2620:113:80c0:8080:10:160:0:207
/usr/bin/iperf3
Connecting to host 2620:113:80c0:8080:10:160:0:207, port 5201
[  5] local 2620:113:80c0:8080:10:160:68:35 port 43226 connected to 2620:113:80c0:8080:10:160:0:207 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-1.00   sec   110 MBytes   927 Mbits/sec    9    223 KBytes       
[  5]   1.00-2.00   sec   108 MBytes   910 Mbits/sec    8    213 KBytes       
[  5]   2.00-3.00   sec   109 MBytes   916 Mbits/sec    2    298 KBytes       
[  5]   3.00-4.00   sec   108 MBytes   909 Mbits/sec    4    286 KBytes       
[  5]   4.00-5.00   sec   110 MBytes   925 Mbits/sec    9    205 KBytes       
[  5]   5.00-6.00   sec   107 MBytes   894 Mbits/sec    9    220 KBytes       
[  5]   6.00-7.00   sec   109 MBytes   915 Mbits/sec    4    159 KBytes       
[  5]   7.00-8.00   sec   110 MBytes   919 Mbits/sec    8    149 KBytes       
[  5]   8.00-9.00   sec   108 MBytes   904 Mbits/sec    5    259 KBytes       
[  5]   9.00-10.00  sec   107 MBytes   900 Mbits/sec    5    216 KBytes       
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-10.00  sec  1.06 GBytes   912 Mbits/sec   63             sender
[  5]   0.00-10.03  sec  1.06 GBytes   907 Mbits/sec                  receiver

iperf Done.

same for iperf3 -6 -c openqa.suse.de. So this looks good so far, same should be applied to all machines, simply over salt seems safe. Done that.

#39 Updated by nicksinger 11 months ago

I've brought back the two power workers (malbec and powerqaworker-qam-1). I see ping fails on the following workers: openqaworker8.suse.de, openqaworker-arm-1.suse.de and openqaworker-arm-2.suse.de which is expected:

openqa:~ # salt -l error -C 'G@roles:worker' cmd.run 'ls -lah /etc/sysctl.d/poo73633_debugging.conf && echo "workaround in place" || true'
openqaworker2.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
openqaworker3.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
openqaworker6.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
openqaworker5.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
openqaworker8.suse.de:
    -rw-r--r-- 1 root root 35 Oct 24 13:27 /etc/sysctl.d/poo73633_debugging.conf
    workaround in place
openqaworker9.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
powerqaworker-qam-1:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
QA-Power8-4-kvm.qa.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
QA-Power8-5-kvm.qa.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
malbec.arch.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
openqaworker10.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
openqaworker13.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
grenache-1.qa.suse.de:
    ls: cannot access '/etc/sysctl.d/poo73633_debugging.conf': No such file or directory
openqaworker-arm-1.suse.de:
    -rw-r--r-- 1 root root 35 Oct 22 19:29 /etc/sysctl.d/poo73633_debugging.conf
    workaround in place
openqaworker-arm-2.suse.de:
    -rw-r--r-- 1 root root 35 Oct 22 19:30 /etc/sysctl.d/poo73633_debugging.conf
    workaround in place

I removed these files now and changed the running value with openqa:~ # salt -l error -C 'G@roles:worker' cmd.run 'sysctl net.ipv6.conf.all.disable_ipv6=0'. I reran the iperf-check and saw >800MB/s for all hosts. The salt change is persisted (in /etc/sysctl.d/99-salt.conf) and also the runtime configuration is set to net.ipv6.conf.$default_interface.accept_ra = 2. I would consider this now as finally done. Any objections? :)

#40 Updated by okurz 11 months ago

  • Status changed from Feedback to Resolved

Thanks, perfect final actions :)

#41 Updated by okurz 11 months ago

  • Related to action #80128: openqaworker-arm-2 fails to download from openqa added

Also available in: Atom PDF