Project

General

Profile

Actions

action #92302

closed

NFS mount var-lib-openqa-share.mount often fails after boot of some workers

Added by okurz over 3 years ago. Updated over 3 years ago.

Status:
Resolved
Priority:
High
Assignee:
Category:
-
Target version:
Start date:
Due date:
2021-06-11
% Done:

0%

Estimated time:

Description

Observation

The nfs mount point systemd unit failed recently (and then turned to ok again) on one of our ARM workers. Likely the problem happens when an ARM machine is rebooted multiple times so that eventually we hit an alert window

Acceptance criteria

  • AC1: No alert about failed systemd services related to NFS mount failing on ARM workers

Suggestions

Rollback

  • DONE: ssh openqaworker-arm-3 "sudo systemctl enable --now salt-minion"
  • DONE: ssh osd "salt-key -y -a openqaworker-arm-3.suse.de && sudo salt 'openqaworker-arm-3*' state.apply"
  • DONE: Unpause alerts for openqaworker-arm-3

Related issues 3 (0 open3 closed)

Related to openQA Infrastructure - action #64941: after every reboot openqaworker7 is missing var-lib-openqa-share.mount , check dependencies of service with openqaworker1Resolvedokurz2020-03-272021-06-11

Actions
Copied from openQA Infrastructure - action #89551: NFS mount fails after boot (reproducible on some OSD workers)Resolvedmkittler2021-03-052021-03-31

Actions
Copied to openQA Infrastructure - action #92969: Failing service os-autoinst-openvswitch after boot of some workersResolvedokurz2021-05-23

Actions
Actions #1

Updated by okurz over 3 years ago

  • Copied from action #89551: NFS mount fails after boot (reproducible on some OSD workers) added
Actions #2

Updated by Xiaojing_liu over 3 years ago

openqaworker-arm-2 can't be reach using ping, so the osd deployment failed on 2021-05-21, but I didn't receive the alerting email.

Actions #3

Updated by okurz over 3 years ago

  • Related to action #64941: after every reboot openqaworker7 is missing var-lib-openqa-share.mount , check dependencies of service with openqaworker1 added
Actions #4

Updated by okurz over 3 years ago

  • Subject changed from NFS mount likely fails after boot of ARM workers to NFS mount var-lib-openqa-share.mount often fails after boot of some workers
  • Priority changed from Normal to High

https://stats.openqa-monitor.qa.suse.de/d/KToPYLEWz/failed-systemd-services shows

Currently failing services
Last update | Host | Failing units | # failed services
2021-05-23 05:48:00 | openqaworker-arm-1 | var-lib-openqa-share.mount, os-autoinst-openvswitch | 2
2021-05-23 03:40:00 | openqaworker13 | var-lib-openqa-share.mount | 1
2021-05-23 03:39:00 | grenache-1 | var-lib-openqa-share.mount, os-autoinst-openvswitch | 2
2021-05-22 04:13:00 | openqaworker-arm-2 | var-lib-openqa-share.mount | 1
2021-05-22 01:47:00 | openqaworker-arm-3 | var-lib-openqa-share.mount, os-autoinst-openvswitch | 2

os-autoinst-openvswitch tracked in #92969

Actions #5

Updated by okurz over 3 years ago

  • Copied to action #92969: Failing service os-autoinst-openvswitch after boot of some workers added
Actions #6

Updated by okurz over 3 years ago

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

@mkittler can you comment about your suggestions in #89551#note-8 . Have you tried them? Should we try again?

Trying on openqaworker-arm-3 with:

cat - > /etc/systemd/system/fix_mounts.service  <<EOF
[Unit]
Description=Fix failed mounts by explicit mount command early (https://progress.opensuse.org/issues/92302)
Before=var-lib-openqa-share.mount
DefaultDependencies=no

[Service]
Type=oneshot
ExecStart=/usr/bin/mount -a

[Install]
WantedBy=multi-user.target
EOF
Actions #7

Updated by okurz over 3 years ago

  • Status changed from In Progress to Feedback
Actions #8

Updated by okurz over 3 years ago

  • Assignee deleted (okurz)

fix_mounts.service fails after boot because it's also started to early:

May 23 19:47:44 openqaworker-arm-3 os-autoinst-openvswitch[2684]: Waiting for IP on bridge 'br1', 247s left ...
May 23 19:47:45 openqaworker-arm-3 mount[1147]: mount.nfs: Resource temporarily unavailable
May 23 19:47:45 openqaworker-arm-3 systemd[1]: fix_mounts.service: Main process exited, code=exited, status=64/n/a
May 23 19:47:45 openqaworker-arm-3 systemd[1]: Failed to start Fix failed mounts by explicit mount command early (https://progress.opensuse.org/issues/92302).
May 23 19:47:45 openqaworker-arm-3 systemd[1]: fix_mounts.service: Unit entered failed state.
May 23 19:47:45 openqaworker-arm-3 systemd[1]: fix_mounts.service: Failed with result 'exit-code'.
May 23 19:47:45 openqaworker-arm-3 systemd[1]: Mounting /var/lib/openqa/share...
May 23 19:47:45 openqaworker-arm-3 systemd[1]: var-lib-openqa-share.mount: Mount process exited, code=exited status=32
May 23 19:47:45 openqaworker-arm-3 systemd[1]: Failed to mount /var/lib/openqa/share.
…

trying to introduce additional dependencies waiting for sockets.target.

Actions #9

Updated by okurz over 3 years ago

  • Assignee set to okurz
Actions #10

Updated by okurz over 3 years ago

  • Description updated (diff)

Tried increased timeout and retries in /etc/fstab

# extended NFS retry https://progress.opensuse.org/issues/89551
openqa.suse.de:/var/lib/openqa/share        /var/lib/openqa/share   nfs ro,x-systemd.mount-timeout=30m,retry=30 0 0

but this does not prevent the mount points to show up as failed for the first five minutes after boot. Also looks like this for the specific mount unit:

# journalctl -f -u var-lib-openqa-share.mount
May 24 14:23:49 openqaworker-arm-3 systemd[1]: Mounting /var/lib/openqa/share...
May 24 14:23:49 openqaworker-arm-3 systemd[1]: var-lib-openqa-share.mount: Mount process exited, code=exited status=32
May 24 14:23:49 openqaworker-arm-3 systemd[1]: Failed to mount /var/lib/openqa/share.
May 24 14:23:49 openqaworker-arm-3 systemd[1]: var-lib-openqa-share.mount: Unit entered failed state.
May 24 14:28:14 openqaworker-arm-3 systemd[1]: Mounting /var/lib/openqa/share...
May 24 14:28:16 openqaworker-arm-3 systemd[1]: Mounted /var/lib/openqa/share.

Trying again with fix_mounts.service on top:

# /etc/systemd/system/fix_mounts.service
[Unit]
Description=Fix failed mounts by explicit mount command early (https://progress.opensuse.org/issues/92302)
After=sockets.target
Wants=sockets.target
Before=var-lib-openqa-share.mount
DefaultDependencies=no

[Service]
Type=oneshot
ExecStart=/usr/bin/mount -a

[Install]
WantedBy=multi-user.target

this both combined seems to work good:

run: 01, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  14:37:26  up   0:15,  1 user,  load average: 1.09, 1.15, 0.93
running
Connection to openqaworker-arm-3 closed by remote host.
run: 02, openqaworker-arm-3: ping .. ok, ssh .. run: 03, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot: "System is booting up. See pam_nologin(8)"
Connection closed by 10.160.0.85 port 22
run: 04, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  14:48:56  up   0:05,  0 users,  load average: 1.41, 1.41, 0.71
starting
run: 05, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  14:51:25  up   0:08,  0 users,  load average: 0.80, 1.17, 0.72
running
Connection to openqaworker-arm-3 closed by remote host.
run: 06, openqaworker-arm-3: ping .. ok, ssh .. run: 07, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  15:03:12  up   0:05,  0 users,  load average: 1.84, 1.47, 0.72
starting
run: 08, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  15:05:40  up   0:08,  0 users,  load average: 0.31, 0.98, 0.65
running
Connection to openqaworker-arm-3 closed by remote host.
run: 09, openqaworker-arm-3: ping .. ok, ssh .. run: 10, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  15:16:48  up   0:05,  0 users,  load average: 2.09, 1.47, 0.69
starting
run: 11, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  15:19:17  up   0:07,  0 users,  load average: 1.22, 1.28, 0.73
running
Connection to openqaworker-arm-3 closed by remote host.
run: 12, openqaworker-arm-3: ping .. ok, ssh .. run: 13, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot: "System is booting up. See pam_nologin(8)"
Connection closed by 10.160.0.85 port 22
run: 14, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  15:30:47  up   0:05,  0 users,  load average: 1.28, 1.33, 0.67
starting
run: 15, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  15:33:17  up   0:08,  0 users,  load average: 0.68, 1.09, 0.68
running
Connection to openqaworker-arm-3 closed by remote host.
run: 16, openqaworker-arm-3: ping .. ok, ssh .. run: 17, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  15:44:22  up   0:05,  0 users,  load average: 1.81, 1.37, 0.66
starting
run: 18, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  15:46:51  up   0:07,  0 users,  load average: 1.04, 1.14, 0.67
running
Connection to openqaworker-arm-3 closed by remote host.
run: 19, openqaworker-arm-3: ping .. ok, ssh .. run: 20, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  15:57:58  up   0:05,  0 users,  load average: 2.06, 1.40, 0.64
starting
run: 21, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  16:00:27  up   0:07,  1 user,  load average: 1.27, 1.32, 0.73
running

so at least 14 successful reboots, the other runs either aborted or accessed the system when it was still in the process of bootup. Tried again with disabled fix_mounts.service and that is not enough:

echo "### disable fix_mounts.service again, only timeout and retry options in fstab"; export host=openqaworker-arm-3; for run in {01..30}; do for host in $host; do echo -n "run: $run, $host: ping .. " && timeout -k 5 600 sh -c "until ping -c30 $host >/dev/null; do :; done" && echo -n "ok, ssh .. " && timeout -k 5 600 sh -c "until nc -z -w 1 $host 22; do :; done" && echo -n "ok, uptime/reboot: " && ssh $host "uptime && sudo systemctl is-system-running && sudo reboot || sudo systemctl --failed --no-legend" && sleep 120 || break; done || break; done
### disable fix_mounts.service again, only timeout and retry options in fstab
run: 01, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  16:21:40  up   0:15,  1 user,  load average: 0.81, 0.84, 0.72
running
Connection to openqaworker-arm-3 closed by remote host.
run: 02, openqaworker-arm-3: ping .. ok, ssh .. run: 03, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  16:33:20  up   0:05,  0 users,  load average: 1.58, 1.48, 0.76
starting
var-lib-openqa-share.mount loaded failed failed /var/lib/openqa/share
run: 04, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  16:35:49  up   0:08,  0 users,  load average: 0.56, 1.13, 0.73
running
Connection to openqaworker-arm-3 closed by remote host.
run: 05, openqaworker-arm-3: ping .. ok, ssh .. run: 06, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  16:46:50  up   0:05,  0 users,  load average: 2.04, 1.50, 0.71
starting
var-lib-openqa-share.mount loaded failed failed /var/lib/openqa/share
run: 07, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  16:49:20  up   0:07,  0 users,  load average: 0.75, 1.15, 0.70
running
Connection to openqaworker-arm-3 closed by remote host.
run: 08, openqaworker-arm-3: ping .. ok, ssh .. run: 09, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  17:00:22  up   0:05,  0 users,  load average: 1.72, 1.36, 0.65
starting
var-lib-openqa-share.mount loaded failed failed /var/lib/openqa/share
run: 10, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  17:02:50  up   0:07,  0 users,  load average: 0.88, 1.18, 0.69
running
Connection to openqaworker-arm-3 closed by remote host.
run: 11, openqaworker-arm-3: ping .. ok, ssh .. run: 12, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  17:13:51  up   0:05,  0 users,  load average: 1.75, 1.39, 0.67
starting
var-lib-openqa-share.mount loaded failed failed /var/lib/openqa/share
run: 13, openqaworker-arm-3: ping .. ok, ssh .. ok, uptime/reboot:  17:16:21  up   0:07,  0 users,  load average: 0.86, 1.19, 0.71
running
Connection to openqaworker-arm-3 closed by remote host.

failing service var-lib-openqa-share.mount. The systemd documentation https://www.freedesktop.org/software/systemd/man/systemd.mount.html# explains "Network mount units automatically acquire After= dependencies on remote-fs-pre.target, network.target and network-online.target, and gain a Before= dependency on remote-fs.target unless nofail mount option is set. Towards the latter a Wants= unit is added as well." which I can confirm. But https://www.freedesktop.org/wiki/Software/systemd/NetworkTarget/ states "By default all remote mounts defined in /etc/fstab pull this service in, in order to make sure the network is up before it is attempted to connect to a network share." which does not seem to work in our case. I suspect an actual bug in wicked. But as I assume no one is willing or able to fix that without us having a small and easy reproducer I am not even bothering reporting that elsewhere.

As alternatives I should try other options in fstab, e.g. nofail,x-systemd.device-timeout=600,noauto,x-systemd.automount

new experiment with fstab entry

# extended NFS retry https://progress.opensuse.org/issues/89551
openqa.suse.de:/var/lib/openqa/share            /var/lib/openqa/share   nfs     ro,,noauto,nofail,retry=30,x-systemd.mount-timeout=30m,x-systemd.device-timeout=10m,x-systemd.automount 0 0 

and no fix_mounts.service.

Running test with

echo -e "### disable fix_mounts.service again, fstab entry noauto,nofail,retry=30,x-systemd.mount-timeout=30m,x-systemd.device-timeout=10m,x-systemd.automount"; export host=openqaworker-arm-3; for run in {01..30}; do for host in $host; do echo -e -n "\nrun: $run, $host: ping .. " && timeout -k 5 600 sh -c "until ping -c30 $host >/dev/null; do :; done" && echo -n "ok, ssh .. " && timeout -k 5 600 sh -c "until nc -z -w 1 $host 22; do :; done" && echo -n "ok, uptime/reboot: " && ssh $host "uptime && sudo systemctl is-system-running && test -d /var/lib/openqa/share/factory/ && sudo reboot || sudo systemctl --failed --no-legend" && echo -n " .. ok" && sleep 120 || break; done || break; done

20 reboots successful, https://gitlab.suse.de/openqa/salt-states-openqa/-/merge_requests/496 ready

Actions #11

Updated by okurz over 3 years ago

  • Description updated (diff)
Actions #12

Updated by okurz over 3 years ago

  • Due date set to 2021-06-11

checking after next weekend(s) if this prevents further alerts.

Actions #13

Updated by okurz over 3 years ago

  • Status changed from Feedback to Resolved
Actions

Also available in: Atom PDF