9fans - fans of the OS Plan 9 from Bell Labs
 help / color / mirror / Atom feed
* [9fans] Ken's file server - insurmountable problem?
@ 2004-01-12 18:55 Lucio De Re
  2004-01-12 23:13 ` Charles Forsyth
  0 siblings, 1 reply; 3+ messages in thread
From: Lucio De Re @ 2004-01-12 18:55 UTC (permalink / raw)
  To: 9fans mailing list

I have to confess that my understanding of the file server is
improving, but evidently not fast enough.  I get:

  squiggle: cp(h0)69.6f[p(h0)19.6-p(h0)94.6] grow from 22406 to 22506 limit 71135 uid=3
  disk cache bucket 8 is full
  fs full uid=3

and I don't really understand how to get around this problem.

The actual definition for the filsys is:

  filsys 2e cp(h0)69.6f[p(h0)19.6p(h0)44.6p(h0)94.6]
  filsys 2edump o

Where the (IDE) disk is a little over 6 Gig, so that we're looking
at a cache of 370 Meg and a fake worm of 1.5 Gig (approximately).

I'm trying to copy around 370 Meg (that's a coincidence, and I'm
sure the actual data will exceed that capacity) to the file system
and the above error occurs.  Forcing a dump, on the obviously
mistaken assumption that it will free cache space (buckets?) did
not help: the error recurred when I tried to copy the remaining
data (the actual bucket number, I think, changed).

If it helps, the amount of data copied  before the failure seems
to be 290 Meg.

++L


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [9fans] Ken's file server - insurmountable problem?
  2004-01-12 18:55 [9fans] Ken's file server - insurmountable problem? Lucio De Re
@ 2004-01-12 23:13 ` Charles Forsyth
  2004-01-13  5:25   ` Lucio De Re
  0 siblings, 1 reply; 3+ messages in thread
From: Charles Forsyth @ 2004-01-12 23:13 UTC (permalink / raw)
  To: 9fans

>>and I don't really understand how to get around this problem.

you'll need to type dump periodically during the copy, to flush the cache.
that often works.
the cache typically needs to be larger than the largest potential file



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [9fans] Ken's file server - insurmountable problem?
  2004-01-12 23:13 ` Charles Forsyth
@ 2004-01-13  5:25   ` Lucio De Re
  0 siblings, 0 replies; 3+ messages in thread
From: Lucio De Re @ 2004-01-13  5:25 UTC (permalink / raw)
  To: 9fans

On Mon, Jan 12, 2004 at 11:13:17PM +0000, Charles Forsyth wrote:
>
> >>and I don't really understand how to get around this problem.
>
> you'll need to type dump periodically during the copy, to flush the cache.
> that often works.

OK, I can perhaps do that, but deleting a whole lot of stuff,
triggering a dump and waiting for it to complete, then trying to
continue the copy (mkfs is clever not to copy unchanged files, so
the deletedd files were restored) didn't make a noticeable difference.

> the cache typically needs to be larger than the largest potential file

Maybe more memory?  I have a puny 32 Meg, maybe I need more of
that? It's a bit of a problem getting more, but I would like to
know if it's worth searching.

++L


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2004-01-13  5:25 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-01-12 18:55 [9fans] Ken's file server - insurmountable problem? Lucio De Re
2004-01-12 23:13 ` Charles Forsyth
2004-01-13  5:25   ` Lucio De Re

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).