Version: 4.8.0-0.nightly-2021-05-13-002125 used hardware Dell PowerEdge R640 Bios version 1.6.13 idrac version: 4.40.10.00 iDRAC 9 Although ironic sets the boot order correctly, booting the image is attempted but fails. The workaround is to boot into lifecycle manager and to attempt OS deploy from it. We see this issue with 4.7 as well.
I wonder if this is another instance of boot manager entries preventing vmedia boot (see https://storyboard.openstack.org/#!/story/2008763 ) Can you check the current boot manager entries and remove any that reference paths on a non existent CDROM (don't remove any generic CD entries)
If possible can you also test with other FW versions? and attach the ironic logs
Looking in the console we see that the boot from ISO is actually attempted, but soon moved to the next one (hard drive).
Created attachment 1784170 [details] boot entries from a setup.
(In reply to Alexander Chuzhoy from comment #3) > Looking in the console we see that the boot from ISO is actually attempted, > but soon moved to the next one (hard drive). Does any error show up while its attempted? (In reply to Alexander Chuzhoy from comment #4) > Created attachment 1784170 [details] > boot entries from a setup. Can you remove those unavailable entries to see if they help
(In reply to Derek Higgins from comment #5) > (In reply to Alexander Chuzhoy from comment #3) > > Looking in the console we see that the boot from ISO is actually attempted, > > but soon moved to the next one (hard drive). > > Does any error show up while its attempted? > > (In reply to Alexander Chuzhoy from comment #4) > > Created attachment 1784170 [details] > > boot entries from a setup. > > Can you remove those unavailable entries to see if they help So removing the unavailable entries does the trick as well. Once all of them are deleted, the nodes boot the image with no issues. Is there a way this automatic deletion could be done with Ironic?
(In reply to Alexander Chuzhoy from comment #6) > So removing the unavailable entries does the trick as well. > Once all of them are deleted, the nodes boot the image with no issues. > Is there a way this automatic deletion could be done with Ironic? It was being discussed here, but no solution yet https://storyboard.openstack.org/#!/story/2008763
we see many failures in cloud.redhat.com production users of assisted-installer, attempt for deployment failed (not even starting ) over issues to boot the .iso
Seems like we should also document this, it looks like the same issue as https://bugzilla.redhat.com/show_bug.cgi?id=2011306 that we just moved to documentation. Derek, do you agree?
(In reply to Bob Fournier from comment #10) > Seems like we should also document this, it looks like the same issue as > https://bugzilla.redhat.com/show_bug.cgi?id=2011306 that we just moved to > documentation. Derek, do you agree? Ya, documenting this is the only option we have to deal with this at the moment.
Push request https://github.com/openshift/openshift-docs/pull/47205. Approved by SME and QE.
Latest version reviewed by SME, QE and peer review squad: https://github.com/openshift/openshift-docs/pull/47205