action #161756
closedQA - coordination #121720: [saga][epic] Migration to QE setup in PRG2+NUE3 while ensuring availability
QA - coordination #123800: [epic] Provide SUSE QE Tools services running in PRG2 aka. Prg CoLo
QA - coordination #137630: [epic] QE (non-openQA) setup in PRG2
IPMI access over IPv6 doesn't work on blackbauhinia size:S
0%
Description
Observation¶
In #153739 blackbauhinia was setup from PRG2e but blackbauhinia-ipmi.qe.prg2.suse.org is only reachable when restricting to IPv4. We should ensure that blackbauhinia-ipmi.qe.prg2.suse.org is reachable without needing to restrict to only IPv4.
Acceptance criteria¶
- AC1: blackbauhinia-ipmi.qe.prg2.suse.org is reachable without needing to restrict to only IPv4
Suggestions¶
- Look into a similar issue #153739 and what we did there
- Either fix IPv6 access with BMC update, config tweaks on https://blackbauhinia-ipmi.qe.prg2.suse.org/ using credentials from https://gitlab.suse.de/openqa/salt-pillars-openqa/-/blob/master/openqa/workerconf.sls or remove the AAAA record from https://gitlab.suse.de/OPS-Service/salt/ if we can't get IPv6 to work
https://imagetester-sp.qe.nue2.suse.org was confirmed to be working using IPv6, confirmed over developer tools, by mkittler and okurz but ipmitool using credentials from https://gitlab.suse.de/openqa/salt-pillars-openqa/-/merge_requests/793/diffs does not seem to work, though IPv4 works.
Suggestions¶
- Try firmware upgrade
- Take a look over https://imagetester-sp.qe.nue2.suse.org if something more needs to be configured to enable IPMI over IPv6
- If not feasible to fix then remove AAAA and reverse PTR from https://gitlab.suse.de/OPS-Service/salt/
Updated by okurz 4 months ago
- Copied from action #159555: IPMI access over IPv6 doesn't work on imagetester - try to update BIOS with physical access size:S added
Updated by okurz 4 months ago
- Related to action #153739: Move of openqa.opensuse.org machine NUE1 to PRG2 - blackbauhinia added
Updated by Julie_CAO 4 months ago
Thank Nan for the update. but we still had not luck, ipmitool did not work, even with "ipmitool mc reset cold".
https://10.145.10.207/tests/14550493
[2024-06-07T07:59:35.727686+02:00] [info] [pid:117557] ::: backend::baseclass::die_handler: Backend process died, backend errors are reported below in the following lines:
ipmitool -I lanplus -H blackbauhinia-ipmi.qe.prg2.suse.org -U root -P [masked] mc guid: Error: Unable to establish IPMI v2 / RMCP+ session at /usr/lib/os-autoinst/backend/ipmi.pm line 45.
@okurz, could you help from your side?
Updated by mkittler 4 months ago · Edited
According to page 150 on https://dl.dell.com/content/manual54104854-benutzerhandbuch-f%C3%BCr-integrated-dell-remote-access-controller-9.pdf?language=de-de IPv6 is supported:
ANMERKUNG: Ab iDRAC-Version 2.30.30.30 unterstützt IPMI das IPv6-Adressprotokoll für Linux-basierte Betriebssysteme.
IPv6 is also enabled in the connectivity settings and also DHCP to get an IPv6 address. Not sure why it isn't working.
EDIT: Looks like disabling IPv6 and enabling it again was a bad idea. Now the management interface is not reachable anymore.
EDIT: It came back. On a closer look it seems that the IPv6 address the machine has is not in accordance with what host tells me. That's probably the problem. I can actually connect via the actual IPv6 just fine (via SSH and ipmitool -I lanplus -H 2a07:de40:b230:1:a330:366a:a0aa:40ab -U root -P '…' sol activate).
EDIT: Not sure why it doesn't take the address from the DHCP server. I played around with the settings but nothing worked. I suppose DHCPv6 needs to be enabled but that was actually enabled from the beginning. Not sure what the relevant DHCP server is. Maybe suttner1.oqa.prg2.suse.org but I couldn't see anything relevant messages in the dhcpd log there.
Updated by mkittler 4 months ago
- Status changed from In Progress to Resolved
I gave up and just configured "our" IPv6 statically. Now everything works - even access using the other previously assigned address which now shows up as "IP-Adresse 2" on the web UI within the IPv6 panel. I think that's good enough.