launchpad crashes when trying to view openteacher 2.x code online

Bug #814632 reported by Arvind S Raj
24
This bug affects 2 people
Affects Status Importance Assigned to Milestone
Bazaar
Confirmed
Medium
Unassigned
Launchpad itself
Fix Released
Critical
Unassigned
loggerhead
Invalid
Critical
Unassigned

Bug Description

Click on "Browse the code" in https://code.launchpad.net/~openteachermaintainers/openteacher/2.x. I tried doing it four times-crashed every time.

The branch appears to be damaged:
'
$ bzr branch lp:openteacher/2.x
bzr: ERROR: No such file: ('information.png-20081222091428-1l7jkfpnqphdx5fa-1', '<email address hidden>')
'

and loggerhead - the code browser - failing is fallout from this.

OOPS-2029CBB101.

KeyError: '<email address hidden>'

    Traceback (most recent call last):
  Module launchpad_loggerhead.app, line 345, in wrapped_app
    app_iter = iter(app(environ, wrapped.start_response))
  Module launchpad_loggerhead.debug, line 106, in wrapped_application
    result = application(environ, response_hook)
  Module __main__, line 165, in wrapped
  Module launchpad_loggerhead.debug, line 45, in wrapped
    return app(environ, start_response)
  Module paste.translogger, line 68, in __call__
    return self.application(environ, replacement_start_response)
  Module paste.deploy.config, line 285, in __call__
    return self.app(environ, start_response)
  Module __main__, line 149, in wrapped
  Module paste.wsgilib, line 179, in catch_errors
    app_iter = application(environ, start_response)
  Module launchpad_loggerhead.session, line 53, in __call__
    return self.cookie_handler(environ, start_response)
  Module paste.auth.cookie, line 305, in __call__
    return self.application(environ, response_hook)
  Module launchpad_loggerhead.session, line 80, in _process
    return self.application(environ, response_hook)
  Module paste.httpexceptions, line 636, in __call__
    return self.application(environ, start_response)
  Module launchpad_loggerhead.app, line 230, in __call__
    return view.app(environ, start_response)
  Module loggerhead.apps.branch, line 180, in app
    return c(environ, start_response)
  Module loggerhead.controllers, line 104, in __call__
    values = self.get_values(path, self.kwargs, headers)
  Module loggerhead.controllers.inventory_ui, line 147, in get_values
    filelist = self.get_filelist(rev_tree.inventory, path, sort_type, revno_url)
  Module loggerhead.controllers.inventory_ui, line 91, in get_filelist
    change=change_dict[revid])
KeyError: '<email address hidden>'

description: updated
Revision history for this message
Deryck Hodge (deryck) wrote :

I can't get at the OOPS info to fill in more info or work out if this is a known issue from other bugs.

Changed in launchpad:
status: New → Triaged
importance: Undecided → Critical
tags: added: oops
Revision history for this message
Arvind S Raj (arvindsraj-deactivatedaccount) wrote :

Deryck,
Do you need another OOPS? Is that the issue? Here's another one anyway just in case. OOPS-2030CBA106.

Changed in loggerhead:
status: New → Triaged
importance: Undecided → Critical
Revision history for this message
Francesco Del Degan (pr0gg3d) wrote :

Hi, I'm trying to figure out what's wrong with that branch.
The first thing is that trying to get that branch locally, bzr starts to sync then fails with

$ bzr branch lp:openteacher/2.x
bzr: ERROR: No such file: ('information.png-20081222091428-1l7jkfpnqphdx5fa-1', '<email address hidden>')

Only the .bzr directory is present, no other files/directories.

a bzr log shows:

------------------------------------------------------------
revno: 143
committer: Marten de Vries <email address hidden>
branch nick: 2.x
timestamp: Fri 2011-07-08 17:00:02 +0200
message:
  Updated index.html file, needs testing
------------------------------------------------------------
.
.
.
.
------------------------------------------------------------
revno: 122
committer: Marten de Vries <email address hidden>
branch nick: 2.x
timestamp: Fri 2011-05-13 16:41:03 +0200
message:
  Removed accidentally added lines and deleted western charfus out of the cyrillic alphabet.

bzr: ERROR: Revision {<email address hidden>} not present in "Graph(StackedParentsProvider(<bzrlib.repository._LazyListJoin object at 0x9a9d8ec>))".

Seems that it lacks of revisions 1-121, and this leads to some tracebacks into loggerhead, complaining about some RevisionNotFound and other kind of exceptions, hence the oops.

In "files" view, a traceback occurs when loggerhead tries to render the "bin" directory, but it doesn't find the associated revision.

  File "/home/pr0gg3d/.bazaar/plugins/loggerhead/loggerhead/controllers/__init__.py", line 104, in __call__
    values = self.get_values(path, self.kwargs, headers)
  File "/home/pr0gg3d/.bazaar/plugins/loggerhead/loggerhead/controllers/inventory_ui.py", line 147, in get_values
    filelist = self.get_filelist(rev_tree.inventory, path, sort_type, revno_url)
  File "/home/pr0gg3d/.bazaar/plugins/loggerhead/loggerhead/controllers/inventory_ui.py", line 91, in get_filelist
    change=change_dict[revid])
KeyError: '<email address hidden>'
----------------------------------------

And in "changes" view, the revisions are re-numbered, where last revision 143 is shown as 25.

Before looking how to catch that conditions (maybe with a more graceful exit), I'm really curious about what have made that branch like it, any ideas?

Revision history for this message
Marten de Vries (marten-de-vries) wrote :

Does anyone know how we can get this branch stable again? It's our main branch for the 2.x series and we use it pretty often. I can just copy the data (we've got plenty of backups) into a new branch, but then we lose all the history which is sometimes useful.

Martin Pool (mbp)
description: updated
Revision history for this message
Robert Collins (lifeless) wrote :

I'm marking the loggerhead task invalid - when dealing with corrupt data like this, oopsing is a valid thing to do :) - the launchpad task needs to stay open on the presumption we'll need to deploy a new bzr asap the cause is fixed in the server [if the server is at fault / can detect and prevent this in principle]

description: updated
description: updated
Changed in loggerhead:
status: Triaged → Invalid
Revision history for this message
Marten de Vries (marten-de-vries) wrote :

'Before looking how to catch that conditions (maybe with a more graceful exit), I'm really curious about what have made that branch like it, any ideas?'

I'm just a user of bzr (so no idea about the internal working), but I can say that the local copies of the branch made before the launchpad version was corrupted (with bzr branch lp:openteacher/2.x) work normally and that we can even use them to pull and push from/to the launchpad master branch. (That's how we're developing now, we send each other emails with a non-corrupt version of the branch if someone doesn't have it... :P). But since it's only the master branch, is it possible that that one is corrupted by e.g. a hard disk failure?

Revision history for this message
Robert Collins (lifeless) wrote :

Have you tried push --overwrite to the branch?

Revision history for this message
Martin Pool (mbp) wrote :

Marten, if you attach a tarball of your working copy of the branch to this bug, perhaps the Launchpad operators can replace the copy on the server with that. I'll save a copy of it so that we can investigate this later.

Revision history for this message
Martin Pool (mbp) wrote :

OK, I can confirm I can branch from lp:openteacher ok, and bzr check on the result passes. I get the 'no such file' error branching from lp:openteacher/2.x. I have a local copy over sftp we can use to investigate the bzr bug (if any).

If push --overwrite does not fix this, you can try:

 - rename the branch to say 2.x-broken
 - push a new branch

Revision history for this message
Martin Pool (mbp) wrote :
Revision history for this message
Martin Pool (mbp) wrote :

apparently-damaged branches ,copied from launchpad

Martin Pool (mbp)
Changed in bzr:
status: New → Confirmed
importance: Undecided → Medium
tags: added: repository stacking
Revision history for this message
Marten de Vries (marten-de-vries) wrote :

> Have you tried push --overwrite to the branch?
Yes, but since the branch was already up-to-date nothing happened.

> If push --overwrite does not fix this, you can try:
>
> - rename the branch to say 2.x-broken
> - push a new branch
I've tried it two times. I once pushed with --overwrite, and once without. Both times it doesn't work, I think because it's still stacked on a (according to the file 814632-branches.tgz damaged) branch: 'Created new stacked branch referring to /+branch-id/450725.' I now reverted it all to the original situation.

> Marten, if you attach a tarball of your working copy of the branch to this bug,
> perhaps the Launchpad operators can replace the copy on the server with that.
> I'll save a copy of it so that we can investigate this later.
Done :)

Thanks everybody for the help!

Revision history for this message
Martin Pool (mbp) wrote : Re: [Bug 814632] Re: launchpad crashes when trying to view openteacher 2.x code online

Hi,

As far as I can tell the trunk is fine. Can't you branch from it?

Revision history for this message
Marten de Vries (marten-de-vries) wrote :

I didn't know that /+branch-id/450725 was the trunk, but indeed, I can branch from the trunk and it's fine.

Still lp:openteacher/2.x is broken, while the local branches work and are ok according to 'bzr check' as far as I know:

marten@marten-pc:~/Programming/Projects/openteacher/2.x$ bzr check
Checking working tree at '/home/marten/Programming/Projects/openteacher/2.x'.
Checking branch at 'file:///home/marten/Programming/Projects/openteacher/2.x/'.
Checking repository at 'file:///home/marten/Programming/Projects/openteacher/2.x/'.
checked repository file:///home/marten/Programming/Projects/openteacher/2.x/ format RepositoryFormat2a()
   211 revisions
   228 file-ids
checked branch file:///home/marten/Programming/Projects/openteacher/2.x/ format Branch format 7
marten@marten-pc:~/Programming/Projects/openteacher/2.x$

At the moment the branch situation is as follows:
- lp:~openteachermaintainers/openteacher/2.x-broken <- the renamed version we used until the problems started
- lp:~openteachermaintainers/openteacher/2.x (lp:openteacher/2.x) <- my local version pushed to the by the rename freed location.
- lp:~openteachermaintainers/openteacher/3.x (was: lp:~openteachermaintainers/openteacher/trunk , is lp:openteacher and lp:openteacher/3.x) <- fine.

Revision history for this message
John A Meinel (jameinel) wrote :

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

...

> At the moment the branch situation is as follows: -
> lp:~openteachermaintainers/openteacher/2.x-broken <- the renamed
> version we used until the problems started -
> lp:~openteachermaintainers/openteacher/2.x (lp:openteacher/2.x) <-
> my local version pushed to the by the rename freed location. -
> lp:~openteachermaintainers/openteacher/3.x (was:
> lp:~openteachermaintainers/openteacher/trunk , is lp:openteacher
> and lp:openteacher/3.x) <- fine.
>

So is your newly pushed-up branch still broken?

The original traceback is indicating that a file is claiming to come
from a revision which isn't in the history of the branch. (Roughly,
I'm not 100% sure on that.)

You're saying that 3.x is fine, I'm wondering if there is just a bad
revision in the history of 2.x that is being processed.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk5VIUEACgkQJdeBCYSNAAPLrQCbBbLw/8PWcSDMKPgkdT94nvhu
BUwAoJRBJ5LC/o7GO6ZAv9qY73SGfetp
=BcUC
-----END PGP SIGNATURE-----

Revision history for this message
Marten de Vries (marten-de-vries) wrote :

'So is your newly pushed-up branch still broken?'
Yes. (e.g. OOPS-ID OOPS-2062CBA489)

'You're saying that 3.x is fine, I'm wondering if there is just a bad
revision in the history of 2.x that is being processed.'

If that's the case, I've no idea how that could've been introduced. We used that revisions (I saw revno: 122 a few posts back) for a pretty long time before they started giving problems.

Revision history for this message
Marten de Vries (marten-de-vries) wrote :

I tried to push the branch to a new location with the --stack-on option. I chose the lp:openteacher branch for that, which has no relation to the 2.x branch (it's completely different code), so I think everything got uploaded from my harddisk (it took longer than it normally does), and now it works :D I updated some names, and now everything works as it did before. Thanks everybody for your help!

William Grant (wgrant)
tags: added: bzr
Changed in bzr:
status: Confirmed → New
status: New → Confirmed
Jelmer Vernooij (jelmer)
tags: added: check-for-breezy
Revision history for this message
Colin Watson (cjwatson) wrote :

I don't know whether this was fixed in the branch itself or whether it was fixed somewhere along the way by porting Launchpad's code hosting to Breezy, but either way I can no longer reproduce this either by visiting https://bazaar.launchpad.net/~openteachermaintainers/openteacher/2.x/files or by branching this directly using bzr, so it seems best to close this. Feel free to reopen it if the same problem still exists somewhere I didn't notice.

Changed in launchpad:
status: Triaged → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Duplicates of this bug

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.