git gc - out of memory

[Simon Strandgaard]

My bare repository is on an OpenBSD machine.
I was unaware of the importance of git gc until today
after investigating a problem with "git pull".
So there hasn't been run git gc on the repository ever.

The biggest file in the repository is a 45 mb file.
The repository size is near 2 gb.

What can I do?

$ git gc
Counting objects: 5934, done.
warning: suboptimal pack - out of memory
fatal: Out of memory, malloc failed8)
error: failed to run repack
$ git --version
git version 1.6.5.2
$ uname -a
OpenBSD amiga.opcoders.com 4.3 GENERIC#698 i386
$ ulimit -a
time(cpu-seconds)    unlimited
file(blocks)         unlimited
coredump(blocks)     unlimited
data(kbytes)         524288
stack(kbytes)        4096
lockedmem(kbytes)    662576
memory(kbytes)       1985524
nofiles(descriptors) 128
processes            64
$ du -ks myrepository.git
1859538 myrepository.git
$



Below is the "git pull" problem I'm having. I think its caused
by the former problem. When pulling it dies because of malloc failure.

prompt> git pull
remote: Counting objects: 280, done.
remote: fatal: Out of memory, malloc failed
error: git upload-pack: git-pack-objects died with error.
fatal: git upload-pack: aborting due to possible repository corruption
on the remote side.
remote: aborting due to possible repository corruption on the remote side.
fatal: protocol error: bad pack header
prompt> git --version
git version 1.6.5.2
prompt> uname -a
Darwin pidgin.local 9.8.0 Darwin Kernel Version 9.8.0: Wed Jul 15
16:55:01 PDT 2009; root:xnu-1228.15.4~1/RELEASE_I386 i386 i386
prompt> sw_vers
ProductName:    Mac OS X
ProductVersion: 10.5.8
BuildVersion:   9L31a
prompt>


Kind regards
Simon Strandgaard - http://gdtoolbox.com/
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to [hidden email]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Alejandro Riveira]

El Sat, 14 Nov 2009 20:26:00 +0100, Simon Strandgaard escribió:

> My bare repository is on an OpenBSD machine. I was unaware of the
> importance of git gc until today after investigating a problem with "git
> pull". So there hasn't been run git gc on the repository ever.
>
> The biggest file in the repository is a 45 mb file. The repository size
> is near 2 gb.
>
> What can I do?

run « git repack -adf --window=memory » on the repo where memory is
escaled apropiately for your machine ?

See « git help repack »

Regards;


--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to [hidden email]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Simon Strandgaard]

On Sun, Nov 15, 2009 at 3:44 PM, Alejandro Riveira <[hidden email]> wrote:
> El Sat, 14 Nov 2009 20:26:00 +0100, Simon Strandgaard escribió:
[snip]
>> What can I do?
>
> run « git repack -adf --window=memory » on the repo where memory is
> escaled apropiately for your machine ?

Thank you Alejandro, it now works!

I think the default is 10, so I tried with window=5 and
it completed a full repack.

$ git repack -adf --window=5
Counting objects: 5934, done.
Compressing objects: 100% (5711/5711), done.
Writing objects: 100% (5934/5934), done.
Total 5934 (delta 3733), reused 0 (delta 0)
Removing duplicate objects: 100% (256/256), done.
$ git gc
Counting objects: 5934, done.
Compressing objects: 100% (1978/1978), done.
Writing objects: 100% (5934/5934), done.
Total 5934 (delta 3733), reused 5934 (delta 3733)
$


Now that it works..
Should I report the original issue as a bug somewhere?
e.g. malloc failed sounds like a bug.



Kind regards
Simon Strandgaard - http://gdtoolbox.com/
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to [hidden email]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Alejandro Riveira]

El Sun, 15 Nov 2009 19:33:27 +0100, Simon Strandgaard escribió:


>>
>> run « git repack -adf --window=memory » on the repo where memory is
>> escaled apropiately for your machine ?
>
> Thank you Alejandro, it now works!

 ell glad it does becouse i mad a typo ...

>
> I think the default is 10, so I tried with window=5 and it completed a
> full repack.

 I was trying to make you use --window-memory=[memory] not --window= ;P

>

>
> Now that it works..
> Should I report the original issue as a bug somewhere? e.g. malloc
> failed sounds like a bug.

 This is the right place. Just wait for someone more knowledgeable than
me ...

>
>
>
> Kind regards
> Simon Strandgaard - http://gdtoolbox.com/


--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to [hidden email]
More majordomo info at  http://vger.kernel.org/majordomo-info.html