The SVN server, when measured independent of ReviewBoard, is fast. I can
check out our entire tree (10GB, 20K files) in a minute. But, I agree that
there might be some other delay between the two servers which is causing
the overall delay.
I have logging enabled. I even enabled log profiling, thinking this might
help me find the problem (although it is also probably slowing down my
system as a whole). What should I be looking for? Something like this?
None - dcarson - /r/46/diff/1/fragment/1052/ - Generating diff file info for
diffset id 83, filediff 1052 took 0.017383 seconds
Also, as far as the cache goes, I do have it configured as follows, but I
have no statistics:
Statistics are not available for this backend.
I thought this was the preferred cache configuration. Do I need something
On Thursday, October 9, 2014 2:55:13 PM UTC-4, chri...@beanbaginc.com wrote:
> Hi David,
> Diff rendering itself is fairly speedy, particularly on a machine like
> you're using. It takes a lot less time than that on some of the ones I've
> seen with 1/4th of the RAM and a single CPU.
> I'd investigate the SVN server. If it's taking a progressively longer
> period of time, perhaps there's something going on there.
> If you enable logging in Review Board and view the diff, you will see
> several entries in the logs detailing the amount of time various
> diff-building operations take. For instance, the amount of time it's taking
> to fetch files from the SVN repository, the amount of time it's taking to
> render the diff, etc. Knowing what these numbers look like will tell me a
> Also, to check, are you using a memcached server, and if so, can you
> confirm it is being used? (There are statistics shown for it in the
> administration UI.)
> Review Board - https://www.reviewboard.org
> Beanbag, Inc. - https://www.beanbaginc.com
> -----Original Message-----
> Date: October 9, 2014 at 11:50:49 AM
> Subject: rendering "diff view" very slow
> > I am setting up 2.0.8 to be deployed very shortly. The only major hurdle
> > is that the switch to the "diff view" for a review request with 10 or
> > files is very slow. I would anticipate users wanting to switch back and
> > forth between "review view" and "diff view" so I must figure this out
> > before deploying.
> > For a typical review with 20 files, each file being perhaps 2K lines
> > and a diff that when downloaded is about 200K bytes, this is what I see:
> > - click on "View Diff" in upper right corner
> > - first file or two renders quickly
> > - 3-10 take a few seconds each
> > - 11-20 get progressively slower, taking 5 or more seconds
> > - switching to diff view always takes this long, not just the first time
> > the "diff view" is rendered
> > This seems excessively slow for the size of the files, the proximity of
> > RB server to the SVN server (same switch!) and the size of the diff. For
> > large diffs, or for cases when it is necessary to switch back and forth
> > between the two views, this becomes almost unusable.
> > Any ideas what I ought to look for?
> > I'm running:
> > - ReviewBoard 2.0.8
> > - CentOS 7.0
> > - older Dell server
> > - 16GB RAM
> > - Intel Xeon 3GHz, 2 CPUs, 4 cores each
> > Thanks.
> > --
> > Get the Review Board Power Pack at http://www.reviewboard.org/powerpack/
> > ---
> > Sign up for Review Board hosting at RBCommons: https://rbcommons.com/
> > ---
> > Happy user? Let us know at http://www.reviewboard.org/users/
> > ---
> > You received this message because you are subscribed to the Google
> Groups "reviewboard"
> > group.
> > To unsubscribe from this group and stop receiving emails from it, send
> > For more options, visit https://groups.google.com/d/optout.
Get the Review Board Power Pack at http://www.reviewboard.org/powerpack/
Sign up for Review Board hosting at RBCommons: https://rbcommons.com/
Happy user? Let us know at http://www.reviewboard.org/users/
You received this message because you are subscribed to the Google Groups
To unsubscribe from this group and stop receiving emails from it, send an email
For more options, visit https://groups.google.com/d/optout.