Looks like it’s fixed in 6.5 Update 1? Although had some response that it does not..
Since the upgrade to ESXi 6.5 on my homelab i got bugged with a strange issue.
All my SATA SSD got really high latency and my VM’s randomly froze when i generated a big file copy actions. Latencies in the 5000/15000ms! Yes IEW!
In the logs on my ESXi host i saw random:
Lost access to volume 583fe5a3-45d215a5-8eb9-00151791b8b6 (SSD01) due to connectivity issues. Recovery attempt is in progress and outcome will be reported shortly.
With ESXI 6.5 the AHCI driver has been updated to a newer version, the native driver, vmw–ahci is used out of the box.
[root@vDrone-ESX01:~] esxcli software vib list | grep ahci
sata–ahci 3.0–22vmw.622.214.171.12464106 VMW VMwareCertified 2016–11–16
vmw–ahci 1.0.0–32vmw.6126.96.36.19964106 VMW VMwareCertified 2016–11–16
So i tried disabling this driver and let it use the sata-ahci driver.
Via SSH: esxcli system module set –enabled=false –module=vmw_ahci
After entering this command, reboot your host.
Using the host client you can check now which driver AHCI is using.
Booted up all my VM’s and presto, no more high latency and freezes. Looks like the new vmw_ahci driver is not that stable yet on some intel chipsets. In my case a X99 chipset.
WOW, Thanks So Very Much for this Blog Post!!! I had been pulling my hair out over the slowness issues on my two ESX hosts over the past week plus. I just made these changes to both hosts and they each displayed more adapters after reboot and performance has been so much better and no more event logs statements about loosing access to vmfs volume! Thanks again!!
No Problemo! Glad it helped 🙂
Thanks alot for the post. I had the same problem on ESXi 6.5 with Samsung SSD.
Glad it helped!
Thanks very mucn for this post.
I get same issue and I think it come from my SSD but not 🙂
After some struggles and consulting Google for longer than I would have liked, this post finally helped me resolve the issues I had with two different SSDs as ESXi datastores.
Thanks for posting this, it’s much appreciated!
Thanks for this, I had the same problem. I disabled vmw_ahci module and rebooted the host and if came up fine, but when I checked the storage adapter was again using the vmw_ahci driver. Should it not be using a different one, and will I run into the same ssd issues again after a while?
The driver should have changed, did you check the command?
Yes, the command was right and it did work but only after I put the host in maintenance mode. after reboot the sata-ahci driver is being used. I’ll monitor for SSD errors for a while, hope they’re gone.
Actually, the errors (when using vmw-ahci driver) eventually caused “loss” of data stores (local and network) and VM freezes but after reboot all was back – until enough errors accumulated to cause the same problem again. So no actual data was lost but it was highly inconvenient.
Thanks for your post.
The same problem here. NUC6. Unfortunately it seems news releases get through with plenty of issues :/
How would you reenable it if new release fixes it?
Via SSH: esxcli system module set —enabled=true —module=“vmw_ahci” 🙂
Thanks a lot mate, it really helped me as my lab environment was constantly freezing due to this problem. I’ve got the problem since last year and almost downgraded my server. Great work!!!!
THANK YOU SOOO MUCH, you have saved me from a couple of weeks worth of stress!
Pingback: ESXI 6.5 在NUC下出现SSD传输缓慢问题 | 新视漫影 :: NVACG :: 【备用端口:5000】
This has fixed the speed issue with my SSD drives but I have an additional 4 port Syba PCIe controller card which is now not detected after disabling vmw_ahci. Has anyone else got an add-in card which is still working after running the above command?
Probably not, the driver of your card is in the VMW_AHCI driver and not in the native.
You could try to make a custom ESXi 6.5 image with the VMW_AHCI driver of ESXi6.0U3. #NoSupport 🙂
Thanks for the reply. Does anyone know if this is logged as a bug report with VMWare as its a pretty fundamental issue to the usability of 6.5. Before finding this post, I have spend at least 3 days trying to fix the issue, replacing hardware and generally going round in circles. They should pull 6.5 as a download until this is fixed.
Yes reported it but the thing is that the problems only seem to happen with non HCL hardware.
Pingback: Intel NUC に ESXi 6.5 を入れる | kurokobo.com
Thanks so much ! You save my day !
By the way, copy/past the command didn’t work for me. I used this one:
esxcli system module set -e=false -m=vmw_ahci
me too n 6.5
I got a gen5 NUC NUC5i5MYHE with a Crucial SSD (MX200 500GB) connected via M.2 and BIOS Mapping it to SATA. Booting ESXi from USB Stick.
After Update to 6.5, I had the same Problems. Slow and buggy. When I tried your solution, it did not find the SSD at all. After a lot try-and-error I found out that the custom NUC Drivers/VIB need to be installed for it to fall back to old ahci.
If you got the Problem, go to: https://vibsdepot.v-front.de/wiki/index.php/Sata-xahci
get he latest drivers/mappings package and install it:
esxcli software vib install -d /vmfs/volumes/… (zip-file)
then disable vmw_ahci and reboot.
Hi, thank you very much for this! You save my day!!
Had also some trouble with the command copied from your post. The right format is:
esxcli system module set –enabled=false –module=vmw_ahci
works like a sharm.
ahh… there is a format issue in the comment field. :-/
It must be a Double hyphen before enabled and module.
Pingback: ESXi 6.5 vmw-ahci から sata-ahci に変更して改善！ | Dark Night Memory
saved me after alread wanted to send back whole server.
HW is actual supermicro serverboard with Intel SSD and Toshiba SSD. After disabing the VMW-ahci driver now all is fine. Really bad bug of the actual ESXI Version!
Thanks again for the post!!
Wolf / Germany
There are some news that Update 1 of ESXi 6.5 (6.5U1) should have solved the issues as it comes with a newer ahci driver according the official release notes. It didn’t in my case and I’ll follow up this posts recommendations later.
By the way. During debugging I figured out, that the disks became much faster with below configuration (from 12’000 to 4’000 m/s latency – which sure is still a lot):
– Change controller to SATA (default SCSI)
– Thick provisoning lazy or full
Oliver / Switzerland
I was a bit too happy to early. The disks became faster, yes but the system itself rather unstable.. It is rebooting every 2 to 3 hours. I’ll roll back.