Bug 1414311 - After failed host-deploy and vdsmd, moving to maintenance hangs [NEEDINFO]
Summary: After failed host-deploy and vdsmd, moving to maintenance hangs
Alias: None
Product: ovirt-engine
Classification: oVirt
Component: Backend.Core
Version: future
Hardware: Unspecified
OS: Unspecified
unspecified vote
Target Milestone: ---
: ---
Assignee: Moti Asayag
QA Contact: meital avital
Depends On:
TreeView+ depends on / blocked
Reported: 2017-01-18 09:41 UTC by Yedidyah Bar David
Modified: 2017-02-20 09:12 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Last Closed: 2017-02-20 09:12:49 UTC
oVirt Team: Infra
mperina: needinfo? (didi)

Attachments (Terms of Use)
engine.log (7.31 MB, text/plain)
2017-01-18 09:41 UTC, Yedidyah Bar David
no flags Details

Description Yedidyah Bar David 2017-01-18 09:41:10 UTC
Created attachment 1242102 [details]

Description of problem:

If host-deploy reinstall fails and leaves vdsmd down, moving the host to Maintenance gets stuck with status 'Preparing for Maintenance'.

A quick search in the code finds HostPreparingForMaintenanceIdleTime, which is set to 300 seconds (also verified that with engine-config), but after more than 5 minutes nothing changed and nothing was written to engine.log.

Restarting the engine solved this. Host was still non-responsive, and moving it to maintenance worked immediately.

Version-Release number of selected component (if applicable):

Current master snapshot

How reproducible:

Not sure, always on my current system.

Steps to Reproduce:
1. deploy a host successfully
2. reinstall it, and make this fail somehow, including vdsmd. Perhaps it's enough to manually stop vdsmd, didn't try.
3. Move the host to Maintenance.

Actual results:

Host is stuck in 'Preparing for Maintenance'.

Expected results:

Host moves immediately, or at most after some timeout.

Additional info:

Restarting the engine solves this.

Comment 1 Martin Perina 2017-02-07 09:43:13 UTC
Didi, is this error reproducable? If not I'd close this as WORKSFORME as we haven't found any clue why it happened and we failed all attempts to reproduce this issue (host was always moved to Maintenance at the end).

Note You need to log in before you can comment on or make changes to this bug.