[OpenAFS] vice partition sizes, and large disks

[Warren.Yenson@morganstanley.com]

Matthew,

At least three points for consideration:

 1. There used to be a contention in the algorithm used in searching for
    free inodes with partitions >20Gb.  I think it's been resolved, but
    our AFS support guru always recommends keeping it smaller than 100Gb

 2. If (when) you have to reboot/restart the fileserver, the partitions
    can be salvaged in parallel.  If you break your raid array into
    n pieces, you may not get n-times faster salvage, but it will be 
    better than waiting for the whole thing in serial.

 3. When fileserver starts, the partitions become available in the order
    they are attached.  So your must-have information should be on /vicepa
    so it is available earlier.

              - Warren

Speaking for me, not for my empoloyer

Warren Yenson          	          mailto: warren.yenson@morganstanley.com
Morgan Stanley                    Tel: +1 (212) 762 2334
750 7th Ave, New York NY 10019    Fax: +1 (212) 762 1009

On Fri, 15 Jun 2001, Matthew Andrews wrote:

> Hello,
>     We are planning on moving to raid arrays(sun T300s) for our afs file
> server disks,
> and I was wondering if there are any special considerations when chosing
> afs partition sizes.
> Does it make sense to use the entire array as a single partition(~500GB)
> or should we break
> it up. what things will be affected by the choice of partition size,
> backups? performance?
> 
> thanks for any information you can share on this subject.
> 
> -Matthew Andrews (mnandrews@lbl.gov)
> 
> _______________________________________________
> OpenAFS-info mailing list
> OpenAFS-info@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-info
>