[RC5] Checkpoint file

Flynn, Steve non Unisys steve.flynn at gb.unisys.com
Wed Oct 24 06:25:35 EDT 2001


You can set the "Run for this long then exit" to 6 hours. The clients will
then shutdown and exit cleanly, writing the partially completed work unit
back into the buff-in file. No need for checkpoint file.

In earlier clients there used to be a checkpoint interval, but it's been
removed. I think it was set to 5 minutes. If that's still the case, then
you're going to be writing about 12k every 5 minutes over the NFS link,
which I wouldn't consider high usage.

> -----Original Message-----
> From: Dennis Lubert [mailto:plasmahh at gmx.net]
> Sent: 24 October 2001 07:23
> To: rc5 at lists.distributed.net
> Subject: [RC5] Checkpoint file
> 
> 
> Hello,
> 
> I have a question to the checkpoint files. How often does the 
> client access 
> them, and can this value be changed ??
> I would like to run my client in an environment, where I have 
> a NFS-mounted 
> home directory and share the same buffer files within them. 
> now I would 
> like to create a checkpoint file for each client, since they 
> can run only 
> 6-12 hours in which they dont finish an ogr block usually. 
> since they share 
> buffer files, the work is lost then... so I want to create a 
> checkpoint 
> file for each, but I may not produce high net traffic, so 
> with mounted nfs 
> home dir, writing to the checkpoint file every 10 secs or so, 
> this would be 
> too much traffic with > 100 clients in a non-switched 10MBit 
> network...
> so how often does it check the files ? or do u use ne other 
> good methods 
> for the client in such an environment ??
> 
> Dennis Lubert
> 
> --
> To unsubscribe, send 'unsubscribe rc5' to 
> majordomo at lists.distributed.net
> rc5-digest subscribers replace rc5 with rc5-digest
> 
--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list