- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Community,
I hope I have placed this request for help in the correct location. I did not see another that was obvious.
I have been trying to install ESXi 8.x on 2 separate Dell Precision T7910 servers without success. They are similarly configured with 128GB Ram and Dual Intel Xeon E5-2697 v3 CPU's.
There is a problem with the ESXi installation in that on both servers I receive a similar error message "TSCs are out of sync: cpu1 > cpu27".
I've spent some time checking the server configurations and moving things around. The only thing I have noticed is when I manipulate the RAM configuration the cpu numbers in the error message change.
The CPU's are identical in both servers. One server has a pair of Intel Xeon E5-2697 v3 CPU's. The other server has a pair of Intel Xeon E5-2690 v3 CPU's.
Is there a workaround for this error or can the installer be customised to bypass this issue.
Please see the attached screenshots showing error message and CPU configuration.
Many thanks in advance.
Michael J
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
-------------------------------------------------------------------------------------------------------------------------------------------------------------
Although I am a VMware employee I contribute to VMware Communities voluntarily (ie. not in any official capacity)
VMware Training & Certification blog
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thanks Scott - much appreciated.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Are you using the Dell customized ISO for ESXi 8.0 that has all the drivers it needs?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
The Dell Precision 7910 is a workstation (not a server) so you won't find it in the VMware HCL for systems and a customised Dell ESXi ISO is unlikely to resolve this TSC issue.
The E5-26xx v3 series of CPUs is also not officially supported under ESXi 8.x either.
https://www.vmware.com/resources/compatibility/detail.php?deviceCategory=cpu&productid=84&deviceCate...
You could try the suggested workaround using the boot option timerForceTSC=TRUE as specified in this KB https://kb.vmware.com/s/article/65186
The E5-26xx v4 (Broadwell) series is supported under ESXi 8.x according to the VMware HCL but it is better to jump this TSC hurdle first before deciding if upgrading to E5-26xx v4 would be worthwhile in your use-case. The E5-26xx v3 and v4 are socket compatible and the Precision 7910 A34 BIOS supports both v3 and v4 CPUs.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Implemented workaround using Ctrl-O and added timerForceTSC=TRUE. Same result on Dell Precision 7910 (2017 vintage). Weird because ESXi 8.0.1 installed without error.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
JFYI - This is a known issue which has already been reported internally to Engineering and a fix has already been implemented and will be available in a future update.
As I understood, depending on how out of sync the TSCs, you may experience different behaviors. One option that was suggested that would at least allow the booting (but subsequent crashes may still occur) is by adding tscSyncSkip=TRUE during boot up (SHIFT+O)
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi, yes I tried using the customized ISO from Dell - VMware-VMvisor-Installer-8.0.0.update02-22380479.x86_64-Dell_Customized-A00. Did not make a difference but looking at the thread responses. Thanks.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
OK, thank you all for the knowledge share and steps I could try.
I tested out the workarounds - both boot options timerForceTSC=TRUE and tscSyncSkip=TRUE.
The results were a little different in that the boot process stated but ended up the same crash. Message is "Requested malloc size 104001552 failed, caller offset 0x453ff17f.
Please see the attached screenshot.
Thank you to all that contributed to finding a solution. Presently running ESXi 7u3 for now as a workaround.
MJ
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
So, I tried the following which I thought would give me the results I wanted but just when I believe ESXi should finally load I see the PSOD. So, I am going to try scripting the settings which may work better. Any guidance?
1. Made sure Bios set to boot from UEFI USB.
2. Using Shift+O to add boot options in the following order seemed to be required: tscSyncSkip=TRUE timerForceTSC=TRUE
3. Now installer for ESXi 8.0.2 working and installed ESXi to an SSD.
4. Once ESXi 8.0.2 installed and running.
5. SSH to console and add.
6. esxcli system settings kernel set --setting=tscSyncSkip --value=TRUE
7. esxcli system settings kernel set --setting=timerForceTSC --value=TRUE
8. Reboot
Just before ESXi is up I receive the PSOD.
Almost there.
MJ
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi All,
1. Made sure Bios set to boot from UEFI USB.
2. Using Shift+O to add boot options in the following order seemed to be required: tscSyncSkip=TRUE timerForceTSC=TRUE
3. Now installer for ESXi 8.0.2 working and installed ESXi to an SSD.
4. Once ESXi 8.0.2 installed and running.
5. SSH to console and add the above settings as shared by the posts to the kernelopt= line. For example my kernelopt finally looks like this -
6. kernelopt=autoPartition=FALSE tscSyncSkip=TRUE timerForceTSC=TRUE
7. Reboot ESXi through the UI.
8. ESXi 8.0.2 came up successfully.
Thank you all for your support.
Made it in the end.
MJ