* The Windows bustage in nsAnnotationService has been solved by booting the
Windows build slaves with /3GB. We got nightlies on all platforms this morning.
* Soon after, I reopened the tree for metered checkins and the first major
checkin broke the tree because of missing files and other errors. That
checkin has been backed out.
* Then, tinderbox.mozilla.org died (disk controller is failing). This is
being tracked in bug 546470, assigned to justdave. The current
tinderbox.mozilla.org is actually tinderbox-stage, and it apparently has
problems that it is not receiving the correct tinderbox mails and is
therefore nonfunctional.
The current tree status is everything is closed, including CVS (because
bonsai.mozilla.org was on the same box as tinderbox.mozilla.org). Once it is
back online, we should still meter checkins very carefully, since we don't
have much performance data on Windows since last Thursday. If nobody
volunteers to do the metering, the tree should remain closed until tomorrow
when there is an active sheriff again.
Happy local hacking?
--BDS
No known fallout at this point. Win32 PGO builds are green, except
where a code change broke it.
> * Soon after, I reopened the tree for metered checkins and the first major
> checkin broke the tree because of missing files and other errors. That
> checkin has been backed out.
>
> * Then, tinderbox.mozilla.org died (disk controller is failing). This is
> being tracked in bug 546470, assigned to justdave. The current
> tinderbox.mozilla.org is actually tinderbox-stage, and it apparently has
> problems that it is not receiving the correct tinderbox mails and is
> therefore nonfunctional.
This problem was resolved by justdave a few hours ago, and tinderbox
is processing mail properly. Some builds will remain unfinished due to
some mail loss, but we have full set of builds on the current tip of m-
c since then.
> The current tree status is everything is closed, including CVS (because
> bonsai.mozilla.org was on the same box as tinderbox.mozilla.org).
justdave says bonsai is back too, so there are no infrastructure
problems remaining.
> Once it is
> back online, we should still meter checkins very carefully, since we don't
> have much performance data on Windows since last Thursday. If nobody
> volunteers to do the metering, the tree should remain closed until tomorrow
> when there is an active sheriff again.
Wanted - sheriff with several hours time to commit to managing tree.
-Nick