Re: lintian 2.5.34 performance results over the first 3 days
On 2015-07-26 13:11, Niels Thykier wrote:
> Hi,
>
> Lintian 2.5.34 has run on lintian.d.o for 3 days (6 runs) now. The
> results are very promising so far.
>
> * It is at 49% reprocessed. At this rate, we can expect it to finish
> after 13 runs (6.5 days).
> [...]
>
Lintian finished last night during the 12th run (6 days), which was 1
run earlier than I estimated. :)
* All of the performance logs for 2.5.34 are now available at [2].
The munin graphs show that:
* We consistently use 150% (of 200%) CPU. This is /mostly/ split in
~100% user and 50% system time.
- Though we do see occasional spikes of +20% (and up to 100%) I/O
wait. This seems to have been worst in the latter part of run 6
and the early part of run 10.
- Load average seems to stay consistently under 3.5.
* Memory usage (still) generally stays under 1.5 GB except for run 3
where it climbed to 2.6GB.
- With this peak, we commit lindsay.debian.org to over 8GB of memory
(with it only having 6GB + 5.5GB swap)
- We seem to be sharing enough memory to avoid swapping considerably,
but if we push it too high, we will start to see failing forks.
- Note that we do *not* seem to be using any particular high amount
of memory in run 10 (where we had high I/O wait), although we seem
to hit a limit for kernel size caching.
I have not checked the log of the latest 6 runs yet. I will file bugs
for the known CPU performance issues (from my previous mail) and any new
ones I might find.
Thanks,
~Niels
> [...]
>
> [2] https://people.debian.org/~nthykier/lintian/performance-2.5.34/
>
Reply to: