[539] in Info-AFS_Redistribution
Re: rs_aix31 server hints?
daemon@ATHENA.MIT.EDU (Bob Andrews)
Thu Jan 16 21:54:19 1992
From: bob@ibmpa.awdpa.ibm.com (Bob Andrews)
To: dstokes@sunlight.austin.ibm.com
Date: Thu, 16 Jan 92 17:40:52 PST
Cc: bob@ibmpa.awdpa.ibm.com, info-afs@transarc.com
In-Reply-To: <9201170133.AA16611@sunlight.austin.ibm.com>; from "dstokes@sunlight.austin.ibm.com" at Jan 16, 92 7:33 pm
dstokes@sunlight.austin.ibm.com writes:
> >up the AFS volumes. If you reserve space by leaving room in the
> >volume group, and then later expand the logical volume because you
> >run into this problem, you can't reclaim the space as reserve space
> >since logical volume can only grow, not shrink.
>
> Yep, thats true. Course I originally saved 30M...:) Thats probably not
> practical for everybody. Wrt to keeping a large file around, I think I
> remember trying only extending the logical volume by 2M to begin with,
> and that wasn't enough. This was a pretty big partition (800M). So if
> you take this approach, should your reserve space be proportional to the
> size of the partition??
>
> Dawn
>
You may be right. Maybe I haven't saved enough space. Usually you're OK
if there are lots of volumes on the disk. The ones that come up OK can
be moved, and then you have space to work with. But that's not quite
good enough to give me a warm-fuzzy feeling about the AFS server. I like
to keep my filesystems < 90% full, and since they are 1.8 Gig, that leaves
a pretty good amount of space (almost 200 MEG) that will have to get used
up before there's a problem.
Anyone at Transarc know if the bug (where the fileserver core dumps when
the /vicep? fs fills up on a rs_aix31 system) is fixed in AFS 3.2?
Bob
