Description of problem: on a ORACLE SERVER X6-2, after writing the image to disk, the server will power off but not completely and won't reboot until we manually reboot it. Introspection worked fine but the deployment always hangs there. Version-Release number of selected component (if applicable): Latest How reproducible: Always Steps to Reproduce: 1. Introspect a node 2. Try deploying on it 3. Actual results: Stuck on reboot after writing to disk Expected results: Reboots and call backs Additional info: Just before rebooting, I see a kernel trace but it does't stick long enough for a screenshot.
Hi David, Would you have an sosreport and/or undercloud logs that would go along with this, especially ramdisk logs? Looking at the attached case its not clear where in the sequence this particular problem happened and which attachments reflect the problem. Thank you.
All I can find is "Host power-on denied because SP is not ready to manage power" in the hardware event logs. I suspect it may be a hardware problem or misconfiguration. As a work around, you can try increasing either or both of https://github.com/openstack/ironic/blob/stable/ocata/etc/ironic/ironic.conf.sample#L1776-L1788. Please let us know if it helps.
Yeah it looks like a power issue or configuration problem in ILOM (SP). From a contact of mine at Oracle, this error message "looks like it happens when the Power State Machine is locked. Which could be due to a progparts update, ueficfg lock, power budget issue, etc." You indicated that you can manually reboot the server. Is that done through the ILOM cli or web interface? If not, can you try to start the host using ILOM and see if the same problem occurs?
Thanks David. I'll close this for now, please reopen if you hear any reports of this happening again.