-
Notifications
You must be signed in to change notification settings - Fork 95
Master build failed #768
Comments
Relevant bits.
|
We need to start capturing the hostd log file as part of the build, the snapshot and restore model for CI also implies that we lose all other logs relevant to the failure. |
@kerneltime My main concern with this failure is that all previous builds of that PR were done successfully on top of master, the error in question came from objtool in vmdk_ops.get_vsan_devfs_path (not in the scope of #760) and the output errors weren't being logged until #765. But from CI history after #760 there isn't any more failures to look at. |
@brunotm I will delete the hostd log from here, we snapshot and restore so it does not contain any relevant bits.. I have filed a issue to capture the hostd logs for CI runs. |
@kerneltime Early finding the #770 was the good part. Thanks. |
@kerneltime While my understanding is that is that the file itself is locked or open by something outside the vmdk_ops service (the busy messages are from failed open() calls on the file) and the ops in parallel tests are being performed on different volumes i don't have the debug log to actually exclude the #760 code change. Even though it seems unlikely that they are the reason for these failures (the locking scope is unchanged) If there was nothing else that we know of happening during that build, i can open a test PR with debugging loglevel hoping to catch something similar, or we can revert and hold back that change and the #770 until this is sorted out. What do you think?
|
Possibly could spawn an lsof during the retry loop to figure which process has the file open |
@govint |
@govint @kerneltime
|
After @brunotm PR was merged #760 the master build failed https://ci.vmware.run/vmware/docker-volume-vsphere/463
ci.463.txt
The text was updated successfully, but these errors were encountered: