Closed Bug 529156 Opened 15 years ago Closed 15 years ago

reclone try server repository

Categories

(mozilla.org Graveyard :: Server Operations, task)

x86
Windows NT
task
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: vlad, Assigned: aravind)

Details

Fetching a URL such as

http://tests.themasta.com/tinderboxpushlog/fetchraw.php?site=pushlog&url=try/pushloghtml%3Fstartdate=12+hours+ago%26enddate=now

which redirects to the pushlog for that tree takes upwards of 45s -- it makes using tbpl on the try server almost unusable.  Any way that we can speed that up?
It's probably time to reclone the try repo - all the heads on it (probably over 1000 at this point) are the source of the slowdown. I'll try to do it along with another downtime this week.
Assignee: nobody → bhearsum
Status: NEW → ASSIGNED
Summary: tbpl fetching pushlog for try server is extremely slow → reclone try server repository
Ah, yeah, that would probably do it.  Any chance we can morph this bug into a way to automate that process, perhaps weekly?  A more complicated option would be to strip the heads (and preceding changesets that don't appear on any other head branches) at the same time as the builds expire from the download site.
(In reply to comment #2)
> Ah, yeah, that would probably do it.  Any chance we can morph this bug into a
> way to automate that process, perhaps weekly?  A more complicated option would
> be to strip the heads (and preceding changesets that don't appear on any other
> head branches) at the same time as the builds expire from the download site.

I just filed bug 529179 for this. I don't think anyone has time to look into it yet, though. We'll do this (again) as a short term fix for now.
Catlee has been working on scheduling this.
Assignee: bhearsum → catlee
(In reply to comment #4)
> Catlee has been working on scheduling this.

Catlee: thanks for waiting for me. I've now got everything I need from the try repo logs, you can reset.

Also, per irc, reassigning to justdave for the 9am downtime.
Assignee: catlee → justdave
Component: Release Engineering → Server Operations
QA Contact: release → mrz
(In reply to comment #5)
> Also, per irc, reassigning to justdave for the 9am downtime.

I don't recall saying I could do this...  be happy to if someone provides instructions, but I have no clue how to do it otherwise...
(In reply to comment #6)
> (In reply to comment #5)
> > Also, per irc, reassigning to justdave for the 9am downtime.
> 
> I don't recall saying I could do this...  be happy to if someone provides
> instructions, but I have no clue how to do it otherwise...

From irc, sorry, this should have been to Aravind.
Assignee: justdave → aravind
Done, per Ben on irc.
Status: ASSIGNED → RESOLVED
Closed: 15 years ago
Resolution: --- → FIXED
I pushed the contents of 1.9.1, 1.9.2, tracemonkey, and electrolysis repositories to try. Places was just recently recloned from mozilla-central, so I didn't push it.

I also had to reconfig the master to fix the HgPoller. It still managed to miss the first push to the newly cloned repo.
Product: mozilla.org → mozilla.org Graveyard
You need to log in before you can comment on or make changes to this bug.