Bug watches aren't getting updated in an ordered fashion

This should be fixed when the bug watch updater checks this bug today (there can be up to a 24 hour delay between a remote bug's status changing and Launchpad pulling in the new status). I can't reproduce the problem locally so we'll wait for the bug watch to be updated in today's run before deciding whether this is actually a problem.

According to the activity log http://bugzilla.gnome.org/show_activity.cgi?id=554628 it was changed on 2008-11-07 20:06:55 UTC so it has certainly been more than 24 hours by now.

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

2008/11/22 era:
> According to the activity log
> http://bugzilla.gnome.org/show_activity.cgi?id=554628 it was changed on
> 2008-11-07 20:06:55 UTC so it has certainly been more than 24 hours by
> now.

According to https://bugs.edge.launchpad.net/bugs/278964/+watch/47324,
the bug watch was last checked on 2008-11-20. Bug watches should be
checked once every twenty-four hours at least, so this is something
that needs looking into.

 status triaged
 importance high
 assignee gmb
 milestone 2.1.12

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: http://getfiregpg.org

iEYEARECAAYFAkkocsEACgkQwutMq02tGPqKZQCcDHHnZiiGLdzbly5PaRQpB6CV
s7wAn1uO89e3n/OLPViUHKmC3ub/iO78
=tAQY
-----END PGP SIGNATURE-----

I think part of the problem is that we're not updating things in a sane order.

Consider this snapshot of the bug watches that need updating for debbugs and gnome bugs, taken at 16:50 UTC using the query at https://pastebin.canonical.com/11566/:

 * debbugs
 -- Total watches: 8728
 -- Watches needing update: 2072
 -- Watches never checked: 232
 * gnome-bugs
 -- Total watches: 9746
 -- Watches needing update: 2418
 -- Watches never checked: 390

Versus this one, taken at ~1800UTC

 * debbugs
 -- Total watches: 8732
 -- Watches needing update: 2078
 -- Watches never checked: 236
 * gnome-bugs
 -- Total watches: 9748
 -- Watches needing update: 2299
 -- Watches never checked: 391

So, unchecked watches are being checked, but there's still a lot of them and considering that we check up to 100 watches every ten minutes there should surely be less (or none) of them over an hour after the initial sample was taken.

ISTM that we're selecting the bug watches to be updated in a haphazard fashion rather than updating the never checked ones first, followed by the oldest ones (though we then have the problem of dealing with bug watches that haven't changed for months, which is the subject of bug 302529). If we start doing this it should clear out the backlog and may fix the issue about which this bug was originally filed.

Didn't land this cycle due to UDS and test failures.

(Dunno if it's at all related, but https://bugs.launchpad.net/bugs/47889 says "Invalid" for the upstream bug when in fact the status for http://bugzilla.gnome.org/show_bug.cgi?id=548211 is "RESOLVED FIXED".)

(Don't know if the following is related as well. Please let me know if I need to create a new bugreport.)

The "CPAN RT" tracker doesn't seem to update as well (or I'm just too impatient?):

https://bugs.launchpad.net/bugs/228679/+watch/50457
https://launchpad.net/bugs/bugtrackers/cpan-rt

2009/1/4 petski:
> (Don't know if the following is related as well. Please let me know if I
> need to create a new bugreport.)
>
> The "CPAN RT" tracker doesn't seem to update as well (or I'm just too
> impatient?):

That looks (from the logs) as though it's a separate problem. Please
file a new bug report at http://bugs.launchpad.net/malone/+filebug and
we'll look into it.

Moving this back to Triaged for now as there's been no significant movement on it this cycle. It's still high priority and needs to be fixed sooner rather than later.

bug 236041 is/was another example

Fixed in devel r7907.

Landed on devel in r7907.

Here's the current unchecked watch counts for debbugs and gnome-bugs (two of the bigger bug trackers):

  name | count
------------+-------
 debbugs | 9453
 gnome-bugs | 10873

We can use this to check that the fix works; it would be a good idea, if it works, to CP it to production as soon as possible.