tech-repository archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

git on small systems

I did some testing over the last little bit for git on small systems.
Please correct anything weird you see; I'm not a git master.

I am testing on my VPC that looks like this:

NetBSD 6.0.2 NetBSD 6.0.2 (PANIX-VC) #0: Sun May
19 03:41:56 EDT 2013

~ $ free -m
        total           used            free
Mem:         487             483               3
Swap:        266             164             102


hw.physmem64 = 536870912
swapctl -l
Device      512-blocks     Used    Avail Capacity  Priority
/dev/xbd1b      546800   337456   209344    62%    0


I was able to do a full git clone from
but it took many hours.
I was getting failures doing git clone from git://
and am still tuning to see if I can get that to work.

In this case git stays around 300-400MB in top.

For even smaller systems, ignore some history:

On a heavily constrained system and for tracking just the branch I'm
running, doing "shallow" clones was actually very fast and
memory-efficient.  Both of examples below report under 40MB used in
top so should be possible on even a 64MB systems with a little swap:

git clone --depth 1 --single-branch -b netbsd_6 git://

^^this is exactly what I would want to do on a tracking system anyway,
so it's a pretty slick option.

Tracking current is even easier:
git clone --depth 1 git://


I just turned these even further down to keep testing the full clone
over git://, but here is my ~/.gitconfig

        windowMemory = 5m
        packSizeLimit = 5m
        deltaCacheSize = 5m
        deltaCacheLimit = 100
        packSizeLimit = 5m
        packedGitWindowSize = 5m
        packedGitLimit = 5m
        deltaBaseCacheLimit = 5m


p.s. the previous values were in the 10m - 64m range but I've lost them

Home | Main Index | Thread Index | Old Index