09-03-2020 06:54 AM
Hello,
Actually I'm having a problem with a UCS C220 M3 server. I have connected management port on dedicated mode, I have connection at the begining but some time later (days or weeks) it lost connection, not http access neither ping.
Some systems partner says me that he have same problem in the past with this UCS model, that it was a recognized bug of Cisco with this server and the only solution was to shutdown servers, disconnect power cable 15 minutes and reconnect. I do that and it exactly works.
But about 1 month later the problem returns to happen and I can't send a technicician every month to a CPD to recover administration. I search for bug but do not find it and I would like to know if I can resolve it upgrading CIMC, changing to shared Loom or changing to a newer server. To any option I need the bug to justify the solution with the customer. Can someone helps me?
Server and firmware versions are:
• UCS C240 M3
• CIMC 3.0
• VMWare 6.5
Thank you,
Oliver R.
09-03-2020 07:26 AM
Suggest you move this over to the UCS/DC community, you're posting in the UC/Collab forum.
09-03-2020 07:40 AM
How can I move it? Have I to create a new one on correct forum?
Thank you,
Oliver R.
09-03-2020 11:27 AM - edited 09-03-2020 11:28 AM
I faced this issue while upgrading the UCS220 M3 to the version 3.X and I have opened a tac and done the same procedure. Its almost a year now so I don't remember the exact version which we got stuck. but the customer is using latest firmware and they haven't reported any issues yet regarding CIMC.
You can check with the UCS experts.
09-03-2020 06:26 PM
There have been a couple sets of CIMC memory leaks, which is what your symptoms sound like.
TAC would normally need a tech support after the OOM condition occurred, in order to see if this is a known existing issue or something new.
Short term, you may want to schedule CIMC reboots (which doesn't reboot the server), every few weeks, to avoid the sneakernet trip, which will reset the CIMC processes and memory usage.
Kirk...
09-04-2020 03:09 AM
Ok, thanks to all replys. I will try to update CIMC firmware, although I think I have installed the last suggested version by Cisco to this UCS, version 3.0(4).
In the same way, I will try to schedule a CIMC reboot, not sure how to do it without shutdown VMWare, I will search about it.
Will inform you with results. In the meantime, any other information will be gratefull.
Sincerely,
Oliver R.
09-04-2020 05:52 AM
09-07-2020 07:47 AM
Thanks for information Kirk, although I am searching how to schedule a reboot of CIMC. I think it can't be done.
I checked installed CIMC firmware is the last release and recommended version (4.0(4r)) to this UCS model (UCS 240 M3). So the only troubleshooting I can do after recover CIMC Management is disable customer's SMTP monitoring or change managment to shared Loom and this way bypass the management port of UCS.
Overmore I read all related bugs, behaviors and caveats of this release but I can not find this issue.
Thank you.
09-07-2020 08:58 AM
Greetings.
You might be able to put in a cron job on a local Linux host that has a task command of:
curl -XPOST -k -u admin:YOURPASSWORD https://x.x.x.x/redfish/v1/Managers/CIMC/Actions/Manager.Reset -d '{"ResetType":"ForceRestart"}'
Kirk...
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide