We had problems with svn commits being very slow on the GDAL
repository. I determined it was because the hook script was invoking
a script to route the update to ciabot and this is apparently not working
properly anymore and it was blocking or locking the repo in some
fashion.
Anyone else expeiencing very long svn commit times (1+ minutes)
might want to check (or have me check) for a ciabot submission item.
Best regards,
--
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up | Frank Warmerdam, warmerdam@pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush | Geospatial Software Developer
On Thu, Aug 23, 2012 at 10:25 PM, Frank Warmerdam <warmerdam@pobox.com> wrote:
Folks,
We had problems with svn commits being very slow on the GDAL
repository. I determined it was because the hook script was invoking
a script to route the update to ciabot and this is apparently not working
properly anymore and it was blocking or locking the repo in some
fashion.
Anyone else expeiencing very long svn commit times (1+ minutes)
might want to check (or have me check) for a ciabot submission item.
I also noted that the post-hook script worked (e.g. GRASS change
reaching the commit mailing list quickly) but then svn is completing
only in 1+ minutes.
I went looking for a related ticket, but overlooked that one as the
title is about *Trac* being slow though it later morphs to being about
svn being slow. The item I identified is specific to SVN.
I also noted that the post-hook script worked (e.g. GRASS change
reaching the commit mailing list quickly) but then svn is completing
only in 1+ minutes.
Would you like me to look into commenting out ciabot requests from the
grass post-commit hook?
Best regards,
--
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up | Frank Warmerdam, warmerdam@pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush | Geospatial Software Developer
"32064 messages since the first one, 6.06 years ago, for an average of
1.66 hours between messages"
I wonder what might slow down this script from one day to the next...
Markus,
OK, well let me know if you come to a decision.
I assume the cia service is acting up but I did not investigate further.
Best regards,
--
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up | Frank Warmerdam, warmerdam@pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush | Geospatial Software Developer
"32064 messages since the first one, 6.06 years ago, for an average of
1.66 hours between messages"
I wonder what might slow down this script from one day to the next...
Markus,
OK, well let me know if you come to a decision.
I assume the cia service is acting up but I did not investigate further.
Frank,
what I don't understand:
in /var/www/svn/repos/grass/hooks/post-commit:
/var/www/svn/hookscripts/ciabot_svn.py "$REPOS" "$REV" "GRASS" &
it is send to bg - still svn is waiting for the job. Is there any way
to launch ciabot_svn.py more independently? At this point the svn
client would return and ciabot_svn.py can finish whenever it wants to.
"32064 messages since the first one, 6.06 years ago, for an average of
1.66 hours between messages"
I wonder what might slow down this script from one day to the next...
Markus,
OK, well let me know if you come to a decision.
I assume the cia service is acting up but I did not investigate further.
Frank,
what I don't understand:
in /var/www/svn/repos/grass/hooks/post-commit:
/var/www/svn/hookscripts/ciabot_svn.py "$REPOS" "$REV" "GRASS" &
it is send to bg - still svn is waiting for the job. Is there any way
to launch ciabot_svn.py more independently? At this point the svn
client would return and ciabot_svn.py can finish whenever it wants to.
Any hint?
Markus,
The intention is that ciabot_svn.py runs independently, but the first thing
it does is run svnlook into a pipe the results of which it reformats and
pipes to ciabot remotely. I *assume* the svnlook is blocking and holding
a lock on the repository while it waits to connect to ciabot. I am also
assuming that the svn submit request tries to get access to the repository
one more time before completion and that is blocked by the svnlook command.
My suggestion, if you want to try, is to have the svnlook output directed
to a file in /tmp and then use that to send back to ciabot so that the
svnlook command finishes quickly.
I glanced at ciabot_svn.py with this idea but decided it was a bit
complicated and not that important to me. You can certainly try.
Best regards,
--
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up | Frank Warmerdam, warmerdam@pobox.com
light and sound - activate the windows | http://home.gdal.org/warmerda
and watch the world go round - Rush | Geospatial Software Developer
On Sun, Oct 7, 2012 at 1:34 PM, Markus Neteler <neteler@osgeo.org> wrote:
On Sun, Oct 7, 2012 at 12:47 PM, Martin Landa <landa.martin@gmail.com> wrote:
Hi,
2012/8/23 Frank Warmerdam <warmerdam@pobox.com>:
OK, well let me know if you come to a decision.
I assume the cia service is acting up but I did not investigate further.
just to know, what is the current status? Commits in GRASS repository
are still taking 1+ minutes which is quite disturbing.
So far I didn't ask FrankW to change anything.
I just discovered:
http://cia.vc/
"We would like to apologise for the slowness of bringing the project back,
the current project manager is ill and finding it hard to work on the project!"
Hence I have commented the post-commit script for CIA.vc and we
are back to speed.
http://cia.vc/
"We would like to apologise for the slowness of bringing the project back,
the current project manager is ill and finding it hard to work on the project!"
Hence I have commented the post-commit script for CIA.vc and we
are back to speed.
http://cia.vc/
“We would like to apologise for the slowness of bringing the project back,
the current project manager is ill and finding it hard to work on the project!”
Hence I have commented the post-commit script for CIA.vc and we
are back to speed.
–
---------------------------------------±-------------------------------------
I set the clouds in motion - turn up | Frank Warmerdam, warmerdam@pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush | Geospatial Software Developer