clean_up_caches needs to better account for image converting
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Ironic |
Fix Released
|
Medium
|
Yuriy Zveryanskyy |
Bug Description
When rebuilding and deploying my overcloud repeatedly on devtest the undercloud has been running out of disk space on /dev/sda1, the images in /mnt/state/
I have been manually deleting them before running up my overcloud again. When /dev/sda1 runs out of space devtest ends up freezing and getting into a unstable state.
To reporduce:
Using devtest:
With the correct env settings and in tripleo-
./devtest.sh # will take a while
source ../undercloudrc
# you can now check /mnt/state/
heat stack-delete overcloud
./devtest_
# let the overcloud try come up again and then check the undercloud master_images directory for cached images.
Running the overcloud up for this 2nd time fails in ironic.
/var/log/
Exit code: 1
Stdout: ''
Stderr: 'qemu-img: error while writing sector 3017344: No space left on device\n'
Changed in ironic: | |
status: | New → Confirmed |
Changed in ironic: | |
assignee: | nobody → Yuriy Zveryanskyy (yzveryanskyy) |
Changed in ironic: | |
milestone: | none → juno-3 |
status: | Fix Committed → Fix Released |
Changed in ironic: | |
milestone: | juno-3 → 2014.2 |
Not clearing images is not a bug - we maintain a local cache. Bug is that we have a bad detection of low disk space. As a workaround you can reduce cache size via option pxe/image_ cache_size.