[p4] Is Sync -f trouble?
jgrills at junctionpoint.com
Wed Nov 1 07:58:41 PST 2006
As far as perforce is concerned, 19000 files isn't that much - I've worked
in places where each branch was 120,000+ files and 4+ gb of file data.
Since you're in the "middle of a release", you probably don't have a lot of
time to be rearchitecting your automated build system. A super easy
temporary fix might be to install a p4 proxy (or two or more) and point the
build machines at the proxies. That will at least move the burden of
transferring the files off your primary server. You might even run a proxy
on each build box, but if you do that it would be best if you can put the
proxy files on a different physical drive from where you sync the files to
so that you're not reading and writing from the same disk and wasting a lot
of time seeking the drive heads around. Depending upon the size of your
network and the way it is configured, it might at very least make sense to
make sure the proxies are sitting on the same switch as the build boxes to
reduce traffic over the backbone of your network.
If all your traffic is over a LAN, you might also make sure that your client
compression settings are turned off so that you're not wasting server CPU
time compressing data.
> -----Original Message-----
> From: perforce-user-bounces at perforce.com
> [mailto:perforce-user-bounces at perforce.com] On Behalf Of
> Venters, Cheryl
> Sent: Tuesday, October 31, 2006 11:15 PM
> To: dlewis78731 at gmail.com
> Cc: perforce-user at perforce.com
> Subject: Re: [p4] Is Sync -f trouble?
> The one thing I should have added that I failed to mention is
> we are running at least 10 builds at one time and all of them
> are running the same way. I'm not sure what is considered a
> lot of files. I know one build we run does a p4 tag first
> then syncs to the subsequent label.
> That label contains 19000 files.
> Our poor performance is usually characterized by slowness.
> Depending on the time of day, it can take three to five
> minutes to simply open P4.
> However, today we were seeing the server actually locked for
> about 2 hours and there were no "unusual" commands happening
> that I'm aware of.
> We have been having a lot of problems and recently moved our
> db files to the local drive versus SAN a few weeks ago. We
> did see performance improvement but that came to a screeching
> halt today. We're also right in the middle of release right
> now and the build load has increased dramatically.
More information about the perforce-user