[OpenAFS-devel] What data from production environments could help OpenAFS development?

[Chaz Chandler]

Yes, the afsio in the codecase does not use the client cache.

Jason Edgecombe <jason=40rampaginggeek.com> wrote:

>On 08/22/2011 10:22 AM, Chas Williams (CONTRACTOR) wrote:
>> In message<4E512186.7020506=40rampaginggeek.com>,Jason Edgecombe writes:
>>> I was wondering. What data from production environments would be helpful
>>> for the OpenAFS developers? Would it help us to set priorities? Would it
>>> help us pinpoint performance bottlenecks? Could the data be gathered
>>> without a significant performance penalty?
>> i think the primary problem here might be that there isnt an 'afs server'
>> benchmark that you can run.  you can bench the drives, network or the
>> cpu individually but there isnt anything (that i know of) to quantify an
>> entire configuration.  collecting sar is nice, but some systems dont have
>> the same capabilities so you might want something that is cross platform
>> (just like afs).
>>
>> talking to the fileservers though would be the best choice.  i gave
>> using afsio some thought but i havent looked into it closely enough to
>> see how much of the client cache is in use (which you would want to test
>> seperately i imagine).
>>
>I thought afsio avoided the client cache. Is there another option to use?
>
>I vaguely recall some idea of recoding client/fileserver operations ans=20
>replaying them to the fileserver. Did anything come of that?
>
>Jason
>_______________________________________________
>OpenAFS-devel mailing list
>OpenAFS-devel=40openafs.org
>https://lists.openafs.org/mailman/listinfo/openafs-devel
>

____________________________________________________________
TRY FREE IM TOOLPACK at http://www.imtoolpack.com/default.aspx?rc=3Dif5
Capture screenshots, upload images, edit and send them to your friends
through IMs, post on Twitter=C2=AE, Facebook=C2=AE, MySpace=E2=84=A2, =
LinkedIn=C2=AE =E2=80=93 FAST=21