[OpenAFS] Re: Server disk operations speed

jukka.tuominen@finndesign.fi jukka.tuominen@finndesign.fi
Mon, 8 Apr 2013 21:24:57 +0300 (EEST)


I'm happy to see these numbers. Eventhough I'm not there quite yet,
they'll keep me trying. And I think they also confirm that it may not yet
be the time for optimization, but rather for troubleshooting.

br, jukka


> On Mon, 8 Apr 2013 19:11:24 +0300 (EEST)
> jukka.tuominen@finndesign.fi wrote:
>
>> thank you for your responses. Before going through them in detail, I
>> would just like to make a reality check. What kind of performance
>> figures should one expect from an averagely working afs network
>> (LAN/WAN)? Say, if you would duplicate roughly the 2000 files/300MB
>> directory in your setup, what kind of rates do you get? That is, is
>> 500-1000KB/s a reasonable starting point for optimization, not a
>> magnitude or two higher?
>
> That's not a transfer rate you need to "just accept"; I'm not sure if
> I've seen a transfer rate that slow to the local machine.
>
> If you want an example from unix... on an underpowered VM with no tuning:
>
> $ find jukka -type f -print | wc -l
> 2000
> $ du -sh jukka
> 305M    jukka
> $ time rsync -a jukka /afs/.localcell/lenny-1/foo3/jukka
>
> real    0m18.777s
> user    0m0.660s
> sys     0m3.244s
> $ time rsync -a 300mfile /afs/.localcell/lenny-1/foo3/300mfile
>
> real    0m11.219s
> user    0m0.416s
> sys     0m3.196s
> $ time rsync -a 300mfile 300mfile.2
>
> real    0m5.435s
> user    0m0.872s
> sys     0m1.424s
>
> That's still pretty slow, but faster than what you mentioned (16 MiB/s,
> 27 MiB/s, and 56 MiB/s). That's also a terrible benchmark since it's
> only one trial and there are a million variables... but if you just
> wanted an example, there is one.
>
> --
> Andrew Deason
> adeason@sinenomine.net
>
> _______________________________________________
> OpenAFS-info mailing list
> OpenAFS-info@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-info
>