bohnsack wrote:I haven't had a chance do any further testing. I am trying to get enough work out of the way to test direct (no switch) client<->server and then with a different switch this weekend. With any luck, I'll have some new data early next week.
I've tested direct connect and things seem just as bad.
Details: WinXP on an HP x4000 workstation with a CAT5e cross-over cable (no switch). Both ends at 100 Mb/s full duplex. As soon as I try moving files via CIFS, the ReadNAS locks up from from a network perspective and a hard reboot is required for it to become pingable again.
I'm going to buy a new GigE switch and CAT5e cables now for testing later today, but I'm afraid I'm going see the same behavior with the new equipment. I'm also planning on using Linux on a Z61m thinkpad for the client, instead of the windows machine.
The ReadyNAS is now in a state where is says it needs to resync, so I'm going to wait to do further tests until that process is complete.
My test plan now looks like this:
a) Wait for re-sync to complete.
b) Connect Z61m Thinkpad running Linux to ReadNAS NV+ via GigE switch and brand new CAT5e cables.
c) Verify iozone operations against local filesystem
d) Run iozone for five minutes against ReadyNAS, followed by a collection and posting of log files. If that works:
- run iozone for 2 hours, collect and post log files
- run iozone for 24 hours, collect and post log files
e) If 5 min, 2 hr, or 24 hr iozone via new GigE switch and CAT5e cables doesn't work, reboot, resync, filesystem check, and try direct connect again (this time with Thinkpad) via new CAT5e crossover cable. Do same iozone tests at 5 min, 2 hr, 24, hour, collecting and posting log files if a failure is seen.
Does this plan seem reasonable? Any thing else I should test? What should I do if things still don't work after all of this?
Thank you.


