No such file or directory error during deployment

[Lucas Alvares Gomes]

Dec 17 15:25:28 localhost.localdomain ironic-conductor[11986]: 2014-12-17 15:25:28.040 11986 ERROR ironic.drivers.modules.iscsi_deploy [-] Deploy failed for instance 8d7252ed-75b9-4114-a4d6-fca98124a22b. Error: [Errno 2] No such file or directory: '/dev/disk/by-path/ip-192.0.2.14:3260-iscsi-iqn-2319f569-73b8-4011-bc93-05a2e27dceea-lun-1-part1'

This seems to come from is_block_device()[1] method from deploy_utils.py when it tried to check if the root_partition is a block device in the work_on_disk() method also in deploy_utils.py:

 def is_block_device(dev):
     """Check whether a device is block or not."""
     s = os.stat(dev)
     return stat.S_ISBLK(s.st_mode)

 def work_on_disk(self ...):
     ....
     if not is_block_device(root_part):
     ....

I believe that after the partitioning we would need to wait until the device and partitions is available, and probably manually call partprobe so that the kernel re-read the partition table of that device.

[Lucas Alvares Gomes (lucasagomes)]

Marking as high because it complete cause the deployment to fail and can happen quite often (~20% of the time)

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/142838

[OpenStack Infra (hudson-openstack)]

Change abandoned by Lucas Alvares Gomes (<email address hidden>) on branch: master
Review: https://review.openstack.org/142838
Reason: Doesn't seem to work

[Dmitry Tantsur (divius)]

Hi Lucas! What's the state of this bug? Could you please unassign yourself, if you're now working on it?

[Lucas Alvares Gomes (lucasagomes)]

Right I will unassigned for now since I don't have much time left to finish it in the moment.

[aeva black (tenbrae)]

Setting back to triaged. Bugs with no assignee can't be "In Progress"

[John Stafford (john-stafford)]

Hi Lucas,

Is this still an issue?

[Lucas Alvares Gomes (lucasagomes)]

Hi John,

I will confirm with the internal qe guys that first found this problem, I believe it still can happen.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/177348

[OpenStack Infra (hudson-openstack)]

Change abandoned by SHIGEMATSU Mitsuhiro (<email address hidden>) on branch: master
Review: https://review.openstack.org/177348

[OpenStack Infra (hudson-openstack)]

Change abandoned by Lucas Alvares Gomes (<email address hidden>) on branch: master
Review: https://review.openstack.org/142838

[Jim Rollenhagen (jim-rollenhagen)]

Is this still a thing?

[Ruby Loo (rloo)]

Not sure what the status is for this. I set it to 'triaged' since it clearly isn't 'in progress'.

[Dmitry Tantsur (divius)]

Looks like nobody was able to reproduce it for a while.