2 Posts
0
687
Dell EqualLogic 6100 with high average latency
I have an array with very high latency mostly on the reads. It's affecting performance on our VMs.
Its 8TB, has solid state drives, iSCSI and has 4, 2TB volumes attached as datastores to vCenter 6.7. I'm not sure where to begin. This is not normal and has recently been discovered. I've added some screen shots.
Any help would be appreciated.
Tekcon76
2 Posts
0
February 24th, 2023 10:00
I did open a ticket and found one of the controllers needed the battery replaced. To do this we needed to move the active ports off of the controller needing the new battery. Moving all the active ports to the same controller seemed to fix the issue.
dwilliam62
1 Rookie
1 Rookie
•
1.5K Posts
0
January 30th, 2023 06:00
Hello,
There is very little information to help you here. What I did notice is that you are using extremely large blocksizes. Large blocksizes provide more throughput but show a greater amount of latency since the ack for the IO takes longer. This alone does not indicate a problem. At peak times you are requesting quite a bit of IO from that member.
Do you have a support contract? If so then open a case. They will gather the diagnostic report from the array, and an Archive from SANHQ.
How many drives are in that 6100? What is the firmware?
I would also make sure VMware is configured to the best practices for EQL SANs.
Common settings like DelayedACK off, MPIO to Round Robin with 3 ios/path, and not sharing VMDKs on single Virtual SCSI controller in each VM. Each VM can have up to four (4) Virtual SCSI adapters.
If you search for EQL SAN TR1091 VMWARE BEST PRACTICES you should find a link to download the PDF.
Regards,
Don
#iworkfordell
dwilliam62
1 Rookie
1 Rookie
•
1.5K Posts
0
February 24th, 2023 10:00
Hello,
Thank you for the update. That makes sense as the cache will go to WRITE THROUGH vs. Write Back.
Regards,
Don
#iworkfordell