Brian Long wrote: >> As I mentioned to James Pearson (but might not've cc'd the list), I am >> beginning to think this may not be an NFS issue at all, but rather an >> i/o problem with RHEL's LVM. > > If you mean the local disk which you are trying to backup, why not try > to dump to another local disk and take the NAS out of the picture? If > /var/tmp is big enough, try dump'ing a small local partition (or > subdirectory) to /var/tmp and see if it's still getting 500KB/sec. > > A way to take dump out of the picture would be to run "dd" from > /dev/zero to your NAS and see what the throughput looks like. For > example, dd if=/dev/zero of=/backups/foo bs=2m count=250. This should > write ~500MB to your NAS and not touch your local disk.
Had a chance to do some testing on filesystem i/o. The very same 150GB dump that took 3+ days via NFS took three hours with the dump directed to local disk (separate, LVM-managed filesystem)--througput was roughly 30X that via NFS. So, looks like we're back to looking at NFS, or possibly the network. As I'd noted previously, similar dumps to the same NAS from a Fedora 13 system did not suffer from this poor througput. -- Tim Evans, TKEvans.com, Inc. | 5 Chestnut Court UNIX System Admin Consulting | Owings Mills, MD 21117 http://www.tkevans.com/ | 443-394-3864 http://www.come-here.com/News/ | [email protected] _______________________________________________ rhelv5-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/rhelv5-list
