Had a crash, Resync Perforamnce Questions

[mooseracing]

I recommend to use the RAM disk test before you make sure that the link is properly saturated.

Read test with same settings as above
I/O per second
11064

MB/s
345

Avg I/O response time
23

Max I/O Response time
110

CPU Utilization
23%

1. Double check that the latest drivers are used

The latest ProWinx64 on Intels site gave me drivers with a date of 4/26/11 version 2.4.36.0, the other server is running 1/7/11

2. doublecheck JF values (sometimes those drop to 1.5k)

Any way outside of the registry to check in Core? They both show 9014

3. If you have more than 4 virtual cores available you can max out the # of workers to be equal to the # of virtual cores.

Not 100% sure, but I have 8 cores, so I should setup 8 workers in I/O meter? That slows things down if I do.

I would also doublecheck the chipset driver to make sure it's up to date. In a few cases this caused general performance issues.
Also please tell us what is the CPU models you're using in the SANs

Single Opteron 6128


Will get back on the chipset.


Thanks

[Max (staff)]

Opteron is not a frequent guest of these forums,
Do you have any news on the Chipset update?

[mooseracing]

I'm not seeing what I would view as a chipset driver available from SM:

Adaptec Storage Manager (ASM)
Link: Download

Description: Matrox G200e Graphics for Windows
Link: Download

Description: AMD SP5100 SATA AHCI Drivers for Windows
Link: Download

Description: AMD SP5100 SATA RAID Drivers (By Adaptec) for Windows

I did update the BIOS while I was in there though. RAM drive test stayed within a couple % or so of the previous. I would expect them to be higher though.

[Max (staff)]

Hey, what timezone are you in?
I think I could remote in to assist you with maxing out the performance

[mooseracing]

Eastern Standard, I'm open for that as long as everything we do keeps at least one of the HA Boxes up, as there are a couple servers on one target.

[Anatoly (staff)]

Please email to support@starwindsoftware.com and Max will contact you as soon as he`ll be opened to do it.

Thank you

[mooseracing]

We did a remote session last Tuesday (IIRC), we picked up a little more speed. Max said he was going to check with some others as something still didn't seem right.

[Anatoly (staff)]

Yes, we are still working onn your issue and trying to figure out what will be the best way for you.

[lohelle]

Any news here? I have a quite simular setup, and I have quite bad sync speed too. Intel/Supermicro 10GbE CX4 dual port (one port used), AMD Chipset (dual quadcore 8389 here).

I can saturate the 10GbE link using iperf if I use 5-6 workers and increase the tcp window size (or use default 8k size and use 30+ sessions (-P 35)

[mooseracing]

Nope, I haven't heard back anything yet, unless I accidentally deleted an email.

[Max (staff)]

Our Q&A department didn't manage to reproduce the issue in their environment.
Will it be comfortable for you If I get our Q&A engineer remote into your environment this week?

[mooseracing]

I don't know if we can this week, our cluster servers started acting real funny so I've been moving alot of resources around and I am trying to get things back under control (ie removing clusters, and going back to an easier environment). Maybe the end of this week or next week.

[Anatoly (staff)]

No problem! We can wait another week