thanks for the heads up. you are amazing!

First, we will work on the update over the weekend, and after that will apply the fine-tuning.

yaroslav (staff) wrote: ↑

Wed Feb 28, 2024 2:46 pm

I will check those in good order.

any update ?

here you go. the logs are reachable on wetransfer link below. I couldn't attach it because the files are bigger than 30MB.

https://we.tl/t-E5so6iUmTA

i'm confused about the strip size / full stripe size configuration.
as per SW recommendations stripe size for SSD / RAID5 configuration is 64kb
are these the right settings?

12/21 21:15:36.116065 b84 Common: CStarWindStorageDevice::hangedTaskExist: Storage device hanged operation: location = E:\SSD Storage\SSD Storage.img, request = 0x0000007DCD0FE2B0, cdb[0] = 0x8A, cdb[1] = 0x00, buf_size = 262144, execution time (13203 msec) more than timeout (7000 msec)! 12/21 21:1...

quick update!! I found some logs that are related to this timeout that is causing this issue StarWind Virtual SAN v8.0.0 (Build 15159, [SwSAN], Win64) Built Jun 2 2023 14:29:14 Windows Server 2012 R2 Standard Edition (v6.3 Build 9600) 12/3 16:01:45.293107 b84 Common: CStarWindStorageDevice::hangedTa...

thank you so much for the clarification. The storage is availabe and it does switch path if iSCSI initiator is "active". Do you see the CSV going offline? NOP, I see only reconnecting message to one CSV out of 3 and this CSV hosts most of the VMS. starwind reconnecting.jpg After rebooting ...

ok, how to solve the timed-out HA device issue? the compute nodes lose connection to CSV1 out of a sudden(reconnecting status on ISCSI initiator) while connected to other CSVs. why is it not switching traffic to CSV1 to the other storage device since it is active-active? we are using two sessions fo...

Hi, I read some recommendations about MPIO settings for compute and storage-separated configurations (NOT HCI) that round robin is better than Least queue depth. Also, I need to learn about the configuration of witness connections from the compute nodes. Our scenario: 3 compute nodes 2 vsan storage ...

Hi,

does the same configuration for witness work for converged infrastructure (3 compute nodes and 2 vsan storage nodes)?

should we apply Least Queue Depth policy on MPIO and MCS policy settings?

Hi,

What is the difference between MCS policy and MPIO policy? both should be Least Queue Depth.

done, does it require any restart to take effect?

I already updated the connection policy on vh01 and vh02. I don't see any iSCSI connection between storage01 and storage02 to update the policy settings.
what do you mean by both nodes? the hypervisor or storage?