Get the driver from Dell. The latest PERC controller is 25.2.0004
This is a critical update. There are issues with 25.2.1.0037. Download SAS-RAID_Firmware_V43G1_WN64_25.2.2-0004_A03.EXE and deploy it through iDrac is the easiest.
Afterward you need the matching async driver.. get the VIB from Vmware
deploy this VIB megaraid-perc9 version 6.901.57.00-1OEM. VMware and dell have instructions
Once you are in iDRAC confirm your SSD drives. Hopefully you have intels and not lite-on drives. If not you will have to do the nautilus update
The firmware shipped will cause extremely slow heavy I/O randomly. Look at the release notes and see if you SSD drives are affected.
I went through this nightmare on 9 servers. Its necessary for stability. Hopefully you only need to update the firmware and controller driver.
Thanks for your detailed response. I'll try it. That said, I have one question, based on conversations I've had with both my Dell Prosupport Engineer as well as my VMware VSAN Sales engineer.
Can you give any sort of details as to the problems with 25.2.1.0037?
I spent a fair amount of time with Dell and VMware, and they both told me that VMware is very strict about testing hardware, documenting it in the HCL and adhering to it. I am trying to avoid VMware support not helping me if they find my configuration doesn't match their HCL exactly. I'm definitely not saying there weren't problems with this version; I'm just trying to avoid having support telling me to upgrade because it doesn't match standards, while at the same time not addressing the core problem I have. While I don't have a problem now, I'm trying to prevent a support issue in the future.
We have Intel SSD drives, so we are fine.
The Dell ESXI 6 customized install doesn't appear to add any addition LSI drivers (see page 6 customization details), so you are likely using the Inbox lsi_mr3 driver that also ships with ESXi 6.0.
I do know there is a bug with VSAN Health where using the Inbox driver gives the warning that your driver is different from HCL and I believe that's probably what you are seeing,I linked the forum discussion here:Re: vSAN Health Plugin HCL Check
If you use firmware 25.2.2.0004, you may want to consider following the non-VSAN VCG and install the corresponding driver VMware Compatibility Guide: I/O Device Search. If you end up using the megaraid_perc9 driver instead of lsi_mr3, you'll also need to run "esxcli system module set --enabled=false --module=lsi_mr3" to force ESXi to ignore lsi_mr3 driver.
If you have no issues with 25.2.1.0037, I'd just stick with it and be happy.
Not sure if this is relevant, but the Dell OEM ESXi image was mentioned. With vSAN it is recommended to avoid OEM images and use the stock image from VMware. Thank you, Zach.
Thanks. I got a similar response from my Dell support engineer. While he didn't say there was anything wrong per se with the Dell ISOs, he did say a lot of what Hogan writes about, and that in the end, the H730 controller firmware is still from Dell; VMware just certifies it. So, Dell would still support us, and even if the Dell support rep (if we were to call in with an issue) were to say "You should upgrade to the latest driver to solve your problem", it wouldn't be the first time that's happened to us. Moreover, in this case, we have a specific reason we couldn't upgrade to the latest.
My Dell guy did say it's an ongoing issue, and they're working on it. From the end of Hogan's article, it also looks like they're at least trying
Not using the H730 for VSAN but just wanted to say thank you for posting this. I couldn't compose more than 4 VMs at a time using View Composer without all guest VMs locking up (R730 with 8 Samsung 850 EVOs in RAID 10 on a H730P). I swapped the lsi_mr3 driver with the megaraid_perc9 one and I see disk pushing to over 1 GBps with under 10 ms latency. I'm never seen such a dramatic effect from a single driver! What a relief!
To be clear, Those drives are not on the HCL, and besides terribly bad consistent IO and low endurance they also lack capacitors to protect the DRAM write buffer. On hard power loss be prepared for silent data corruption.
Hi we are running VSAN with this card and we use Dell driver located in this site ESXi 5.5 and vSphere 6.0 stability or performance issues when using PERC 9 | Dell US
install this driver if you want to use Pass-Through for the disk.
We had a discussion with VMware and Dell engineer regarding the H730 and H730 mini card and we ware the first company in Sweden that orded this cards and R730 servers from Dell. so we have tried everything that you could try with this card.
hope it helps
This newer KB article published by VMWare suggests the driver/firmware combo below in addition to updating backplane firmware:
Driver: megaraid_perc9 version 6.902.73.00-1OEM
Controller firmware: 25.3.0.0016
Driver: lsi-mr3 version 6.606.12.00-1OEM
Controller firmware: 25.3.0.0016
In addition to this, if you have Intel flash drives (Dell re branded count here) they have pushed out a new SSD firmware. They had an issue with bad error codes it looks like from the Dell patch notes, and this could cause the H730 to drop the flash drive.