[OpenAFS] Server disk operations speed

[Jukka Tuominen]

Again numbers I'd be very happy with. Thank you for sharing them!

To simplify the problem, I will first concentrate on client/server issues
on the server machine only (appreciating the comments that the actual use
case needs to be optimized on a separate client eventually). This approach
also allows me to run the GA optimizer dealing with both client-side and
server-side parameters, or any parameters that can be reset dynamically
between test cycles, for that matter. BTW, I couldn't extent the token
lifetime from the current 10 hours, eventhough I thought I found the place
to redefine that. It would ease testing if anybody could advice on this.

I think I will next start chasing the data transfer path, and secondly
figure out how to reduce the latency should it be an issue in my setup.

Any new suggestions are highly appreciated.

br, jukka


> Subject: Re: [OpenAFS] Server disk operations speed Date: Mon, Apr 08,
> 2013 at 07:11:24PM +0300 Quoting jukka.tuominen@finndesign.fi
> (jukka.tuominen@finndesign.fi):
>>
>> Hi all,
>>
>> 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?
>
> GE LAN, client and server connected in same switch. 1.6.2 server, 1.4
> series client.
>
> Copy inside AFS, same volume, same file server.
>
> $ du -sh ../infected/;  time cp -r ../infected  .
> 66M     ../infected/
>
> real    0m17.009s
> user    0m0.000s
> sys     0m8.457s
> $
>
> 3.8 MB/sec.
>
> Copy out of AFS, to an iSCSI volume also on local LAN.
>
> $ du -sh ../infected/;  time cp -r ../infected /scratch/mansaxel/
> 66M     ../infected/
>
> real    0m0.546s
> user    0m0.004s
> sys     0m0.416s
>
> Hot cache. 120+MB/sec.
>
> Into AFS from local file system:
>
> jaja.besserwisser.org$  time cp -r  /scratch/mansaxel/infected .
>
> real    0m9.113s
> user    0m0.000s
> sys     0m9.077s
>
> 7.24 MB / sec. Also hot cache, so probably close to max for this test
> case.
>
> All of this is purely anecdotal and not very controlled environment,
> but pointers to ballpark figures it is.
>
> --
> Måns Nilsson     primary/secondary/besserwisser/machina
> MN-1334-RIPE                             +46 705 989668
> You mean you don't want to watch WRESTLING from ATLANTA?
>