Help to debug openssh import failure

Asked by Colin Watson on 2010-03-21

I'd like to try to figure out what's going on with the Launchpad code import of openssh (https://code.launchpad.net/+branch/openssh), which is failing to import any new revisions despite there being plenty of new stuff in CVS. Would it be possible to get a temporary change applied to crank up debugging levels on the code import machines? Or is it worth just cranking up debugging levels across the board?

The change I'm thinking of (albeit untested - I'm not set up to run the Launchpad tests here and would appreciate help) would be something like this:

  http://paste.ubuntu.com/397948/

Question information

Language:
English Edit question
Status:
Solved
For:
Launchpad itself Edit question
Assignee:
Michael Hudson-Doyle Edit question
Solved by:
Michael Hudson-Doyle
Solved:
2010-04-07
Last query:
2010-04-07
Last reply:
2010-04-07

The import failed again, like this: http://people.canonical.com/~mwh/vcs-imports-openssh-main.log (warning: 12 megs)

Is this helpful?

Colin Watson (cjwatson) said : #2

Not really. That's the transient failure to talk to the CVS server, which may be worth debugging at some point but wasn't what I was trying to debug. What I need is the debug log of an import that claims to be successful, but in fact doesn't import any new revisions.

Colin Watson (cjwatson) said : #4

Waitaminute.

totla running with ['--strict', '-b', '/srv/importd.staging.launchpad.net/data/worker-for-branch-29296/bzr_working_tree', '-SC', u'MAIN.6032::', '/srv/importd.staging.launchpad.net/data/worker-for-branch-29296/bzr_working_tree']

Why is LP telling cscvs to start the import at revision 6032? The published lp:~vcs-imports/openssh/main only goes up to revision 5963 (and the cscvs-id property of the tip revision is MAIN.5963)! This would explain the problem; could you look at the bzr working tree it's using and see if it's ahead of what's published? It very much looks like it must be.

Is the import behaving itself now?

Colin Watson (cjwatson) said : #6

Everything seems fine now. Thanks for all your help!

Colin Watson (cjwatson) said : #7

Of course there've been no CVS commits since we repaired it, anyway - but I'll keep an eye on it and file a new ticket if necessary.