550256 - Reset try repo again

Status: RESOLVED FIXED

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

Description

[Jeff Muizelaar [:jrmuizel]]

Requests like http://hg.mozilla.org/try/rev/94ca15b6eebb can take up to 1m and 8s to service. We should be able to better than that.

Comment 1

[Chris AtLee [:catlee]]

This is probably because try ends up accumulating hundreds of heads.  Maybe a sign we need to reset the try repo again?

Does mozilla-central suffer from the same problem?

Comment 2

[Jeff Muizelaar [:jrmuizel]]

It is better, but still quite slow:

time wget http://hg.mozilla.org/mozilla-central/rev/7e39e554e321

--15:13:22--  http://hg.mozilla.org/mozilla-central/rev/7e39e554e321
           => `7e39e554e321'
Resolving hg.mozilla.org... 63.245.208.189, 63.245.208.188
Connecting to hg.mozilla.org|63.245.208.189|:80... connected.
HTTP request sent, awaiting response... 200 Script output follows
Length: unspecified [text/html]

    [  <=>                                ] 7,693          8.18K/s             

15:13:27 (8.17 KB/s) - `7e39e554e321' saved [7693]


real	0m5.051s
user	0m0.002s
sys	0m0.005s

Comment 3

[Nick Thomas [:nthomas] (UTC+12)]

(In reply to comment #2)
> It is better, but still quite slow:
> time wget http://hg.mozilla.org/mozilla-central/rev/7e39e554e321
> real    0m5.051s

Bug 550280 for that.

Comment 4

[Nick Thomas [:nthomas] (UTC+12)]

AFAICT the try repo was last reset in bug 529156 at the start of December.

Comment 5

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

I'll grab this for now - once I extract historic info from try server repo, I'll kick this over to IT for delete/reset of the try repo. 

jeff: if you see delays happening with other repos, please file separate bugs, ok?

Comment 6

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

data extracted for march-so-far. 

aravind: thanks for waiting, its now ok to reset this try repo.

Comment 7

[Chris AtLee [:catlee]]

Happening March 12th, 2010 at 7pm PST.

Comment 8

[Aki Sasaki (not active)]

How do I get the m-c and m-1.9.2 data in here? Check out Try, pull from m-c and commit/push?

Comment 9

[[e-mail] (:bhearsum)]

(In reply to comment #8)
> How do I get the m-c and m-1.9.2 data in here? Check out Try, pull from m-c and
> commit/push?

or clone m-c/1.9.2/1.9.1 and then push to try from that dir.

Comment 10

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

data re-extracted for march-so-far.

Comment 11

[Aki Sasaki (not active)]

It's a good thing I pushed 1.9.1 and 1.9.2:

akira-sasakis-macbook:mozilla-1.9.1 asasaki$ hg push -f ssh://hg.mozilla.org/trypushing to ssh://hg.mozilla.org/try
searching for changes
remote: adding changesets
remote: adding manifests
remote: adding file changes
remote: added 4702 changesets with 8530 changes to 7032 files (+25 heads)
akira-sasakis-macbook:mozilla-1.9.1 asasaki$ cd ../mozilla-1.9.2
akira-sasakis-macbook:mozilla-1.9.2 asasaki$ hg push -f ssh://hg.mozilla.org/trypushing to ssh://hg.mozilla.org/try
searching for changes
remote: adding changesets
remote: adding manifests
remote: adding file changes
remote: added 2297 changesets with 7070 changes to 8007 files (+13 heads)

Comment 12

[Jeff Muizelaar [:jrmuizel]]

This is really bad again.

time wget http://hg.mozilla.org/try/rev/9f3391b5ca0c
--10:57:28--  http://hg.mozilla.org/try/rev/9f3391b5ca0c
           => `9f3391b5ca0c'
Resolving hg.mozilla.org... 63.245.208.188, 63.245.208.189
Connecting to hg.mozilla.org|63.245.208.188|:80... connected.
HTTP request sent, awaiting response... 200 Script output follows
Length: unspecified [text/html]

    [            <=>                      ] 875,307      291.45K/s             

10:58:03 (290.88 KB/s) - `9f3391b5ca0c' saved [875307]


real	0m35.306s
user	0m0.009s
sys	0m0.043s

Comment 13

[Armen [:armenzg]]

Is there a way we could automate resetting the try repo?

For joduinn's stats we could also automate getting the stats before deletion.

Comment 14

[Chris AtLee [:catlee]]

I'd like to know if we could automate pruning old heads from the try repo.

Comment 15

[Chris AtLee [:catlee]]

(In reply to comment #14)
> I'd like to know if we could automate pruning old heads from the try repo.

Filed bug 554656 on this.

Comment 16

[Aravind Gottipati [:aravind]]

Automating the try repo reset is certainly possible (at least on the IT side).  This bug was opened to reset the try repo for one particular instance, and that was done.  If this nesds to happen again, it has to be co-ordinated with build and release and they need to open a bug when they are ready.  Re-opening bugs that are done is not the way to go.