On Thu, Jun 28, 2012 at 9:57 AM, John McEntee <[email protected]> wrote: > Ok, and update on my progress as I have not fixed it yet. > [...] > > I also have 2 scripts running every minute via cron (from a linux box). One > copies a 10 MB file via cifs the other via nfs, neither have been delayed > for more than 5 seconds, normally completes with 1 second. These are done to > their own zfs file systems at the moment, but I can foresee having to point > them at the production filesystem if it does not detect the freezes that the > users report. > > John
You'll need to gather some data when the slowdown happens. I would start at the network with a snoop capture that shows both the clients involved and the AD server in use. There are some cases where the SMB server currently needs to go talk to an AD server to find out the name for the account represented by some SID, etc. and that can cause temporary pauses like you see. If it's anything like that, the network trace will show it happening. -- Gordon Ross <[email protected]> Nexenta Systems, Inc. www.nexenta.com Enterprise class storage for everyone _______________________________________________ OpenIndiana-discuss mailing list [email protected] http://openindiana.org/mailman/listinfo/openindiana-discuss
