Created attachment 1051917 [details] ironic-conductor Description of problem: One of my nodes is reporting the below : ironic-conductor.log:2015-07-14 07:33:24.953 24472 ERROR ironic.drivers.modules.deploy_utils [-] Command: sudo ironic-rootwrap /etc/ironic/rootwrap.conf parted -a optimal -s /dev/disk/by-path/ip-192.0.2.23:3260-iscsi-iqn-ee4db5c0-985b-4a1e-a74b-8210ce9e2673-lun-1 -- unit MiB mklabel msdos mkpart primary 1 2 mkpart primary 2 40962 set 2 boot on When attempting to deploy any overcloud, the node that experienced the ERROR above never is utilized.
ironic logs : http://perf1.perf.lab.eng.bos.redhat.com/jtaleric/BZ/ospd/ironic-conductor.tar.gz nova logs : http://perf1.perf.lab.eng.bos.redhat.com/jtaleric/BZ/ospd/nova-logs.tar.gz
What caught my eye is *dozens* of the following errors every 2 seconds: NodeAssociated: Node 0a9965c0-d483-49f7-ae93-22a09de7e4dd is associated with instance 6027bafb-d1bc-46c3-9646-2abdd93134a0. Looks like nova tries to schedule on the same occupied node... and partitioning failures appear only in the beginning, not sure they're actually root cause of the problem.
Hi Joe, Strange error indeed, it's like the disk was being accessed by another process at the moment Ironic was trying to create the partitions. Is this bug reproducible somehow?
Hey Lucas - I haven't seen this error in awhile. If it reproduces I will be sure to update this bz. Do the logs not provide enough insight?
(In reply to Joe Talerico from comment #6) > Hey Lucas - I haven't seen this error in awhile. If it reproduces I will be > sure to update this bz. Do the logs not provide enough insight? Yeah I can't figure out the reason just by looking at that error. I will mark this as closed but feel free to re-open in case you hit this problem again. Thanks for reporting