[nycbug-talk] Git Tangent: [Was: svnup(1) - worthy of promotion to base?]

[Fabian Keil]

"Isaac (.ike) Levy" <ike at blackskyresearch.net> wrote:

> On Mar 11, 2013, at 3:09 PM, Fabian Keil wrote:
> > "Isaac (.ike) Levy" <ike at blackskyresearch.net> wrote:
> >> On Mar 11, 2013, at 11:31 AM, Fabian Keil wrote:
> >>> Git is really quite impressive on all these fronts- to bad it's getting so obese in the feature implementation department (build it from source one day, you'll see what I mean :)
> > 
> > Without the svn support most of the obesity seems to be gone.
> > And building everything including clang and Firefox from source
> > helps to keeps things in perspective …
> 
> Well, perhaps clang is large- but I don't believe it requires both python and perl to build 'feature-expect complete' :)

I guess it depends on your feature expectations.

The "stripped-down" clang in base doesn't, but the port
depends on both python and perl.

> >> Just curious- (mostly because I'm amazed to see how much more feasable using ZFS has become with smaller memory footprints).
> > 
> > After a bit of tuning it always worked for me.
> 
> That's pretty awesome- do you have, or could you recommend some notes?

On this laptop I merely set:

vm.kmem_size="3G"
vfs.zfs.write_limit_override=300000000
vfs.zfs.arc_max="1200M"
vfs.zfs.arc_meta_limit="600M"

With the exception of the last one they are all mentioned
in both the handbook and the Wiki.

I still occasionally experiment with other values and sysctls
and wouldn't recommend applying any of them without testing,
but they (mostly) work for me and the remaining issues aren't
bad enough to consider going back to UFS.

> I've tried ZFS on boxes with <4gb memory, and failed miserably at making them run reasonably...

Are you referring to reliability issues or "just" performance?

Fabian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
URL: <https://lists.nycbug.org:8443/pipermail/talk/attachments/20130313/7e6115bc/attachment.bin>