[OpenAFS] WinSCP onto debian server/openafs storage errors...

Lars Schimmer l.schimmer@cgv.tugraz.at
Tue, 15 May 2007 16:20:05 +0200

Hash: SHA1

Jeffrey Altman wrote:
> Lars Schimmer wrote:
>>>> This method works sometimes, sometimes not. The connections breaks d=
>>>> nearly everytime while transferring bigger files.=20
>>> Why is the connection failing?
>> Some more tests.
>> 1. WinSCP direct onto local HD of debian sarge system - no error at al=
>> 100% in 1 try
>> 2. WinSCP onto debian sarge into openafs space - round about 7% of 8 G=
>> went through, error, retry & resume - most times 100% of file went int=
>> AFS space after x retries - OpenAFS 1.4.2
>> 3. Like point 2, but Debian sid amd64 system, OpenAFS 1.4.4 and it
>> doesn=B4t resume, restart every retry at 0% - went on forever....
>> Error messages from trying out no. 3 (same errors like on no.2):
> Is the AFS client cache large enough to hold the files?

No, it is not. Cache is about 1-2 GB in size.

> If not, when the cache fills the client has to find and flush
> dirty buffers to the file server before the file is closed. The problem
> is with the scp client not waiting for space to be made to receive more
> of the file.  You might want to turn off storebehind mode so that dirty
> buffers are written to the file server as they are received.

Ok, til yet fs storebehind was not known for me.
And the description is kinda ...
It tells me:
fs storebehind
Default store asynchrony is 0 kbytes.

Does this mean, storebehind is off?
If yes, OK, I see I need a 16 GB cache for OpenAFS on that machine or
copy that file via scp to local home on debian server and afterwards
moving into afs space.

Lars Schimmer
- --
- -------------------------------------------------------------
TU Graz, Institut f=FCr ComputerGraphik & WissensVisualisierung
Tel: +43 316 873-5405       E-Mail: l.schimmer@cgv.tugraz.at
Fax: +43 316 873-5402       PGP-Key-ID: 0x4A9B1723
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org