Something you may want to check is if the server is currently up to date on the BIOS, iDrac, H310 firmware, as well as the drives. I would verify that it is up to date and current and see if anything is then reported, or if the issue resolves itself due to the update. What are the current versions you have? I can get you the links to the ones you need. Are you getting any amber lights on the server or drives?
This server was deployed to the production about 2 weeks ago and everything was updated to the latest version. So unless there have been a firmware or driver update that came out in the last 2 weeks, everything should be of the latest.
the host machine have no antivirus on it, no roles nor any other configuration. just with hyperv enabled and vss disk queue disabled. all client related services are run from vm. Since reboot with all vm off does not eliminate the performance issue, only a power off did the trick, I started to think there is something up in the idrac or firmware or how p310 handles raid 1 with hotspare.
We have a server that is doing the exact same thing at a clients site. PE R420. Raid 1. Server 2012 R2 used as a terminal server. This morning about 4 minutes after a user logged in the entire server started running slow. A reboot didn't resolve the issue. I'm currently running a system state restore to a earlier date but now that i've seen this while its running i wish i could try the cold boot option before this completes....guess i'll find out shortly....
Yes i have seen this thread. The server in question also isnt running any VM's or even has the hyper-v service installed. Also the server is using a Perc H710 mini controller.
The issue may well be with using a H310 controller in this enviornment. The controller does not have write cache and therefore not designed for environments with Virtual Machines. Check your disk queue lengths to verify whether this is your issue. You may want to check out the following thread.
The disk queue's are written before machine is shut down and therefore clear on the next reboot. Again, the H310 is not meant for an environment with Virtual Machines as it has not write disk cache. You may want to GOOGLE H310 + Performance + HYPER V before ruling this out as the issue.
Think about it, if the H310 is really the issue, the symptom will be persistant.
Ever since that time it had a slow down and I hard powered it off, I have yet to re-encounter the issue. Currently i have 2 VM running on top of it over 2012 R2 standard with Hyper-V, no issues.
The best I can explain it is a firmware issue or something with disk resyncing.
Just so you are all aware, my issue has so far been fixed by going in to the BIOS (V2.2.0) and setting the CPU Performance to "Performance" instead of "Performance Per Watt (DAPC). I found that the CPU clock was limited inside of windows to 150mhz (0.15Ghz) which is why the server was running so slow. It would never increase (decrease) its throttling when the server was under heavy load. (surprising windows could run with only 150mhz). A picture of the CPU in action.
I realize this has been stale for awhile, but it came up in my search for an answer to why my R420 VMware host's CPU was not performing well all-of-the-sudden for its two guests. An esxtop command showed very high PCPU UTIL but very low PCPU USED, meaning the CPUs were working really hard but not getting much done.
I see in April Dell released a BIOS update 2.4.2 for the R420. I'm on 2.3.3 so will plan to update ASAP.
"Updated Intel Management Engine (ME) binary to Rev 2.1.5. 0x8B, to address issues that resulted in unexpected temporary or permanent Processor frequency degradation."
I'll also note that a warm restart didn't help, but I'm hoping a cold boot / BIOS update will.
DELL-Chris H
Moderator
•
9.7K Posts
0
September 11th, 2014 13:00
Cirlare,
Something you may want to check is if the server is currently up to date on the BIOS, iDrac, H310 firmware, as well as the drives. I would verify that it is up to date and current and see if anything is then reported, or if the issue resolves itself due to the update. What are the current versions you have? I can get you the links to the ones you need. Are you getting any amber lights on the server or drives?
Let me know what you see.
cirlare
10 Posts
0
September 12th, 2014 10:00
This server was deployed to the production about 2 weeks ago and everything was updated to the latest version. So unless there have been a firmware or driver update that came out in the last 2 weeks, everything should be of the latest.
the host machine have no antivirus on it, no roles nor any other configuration. just with hyperv enabled and vss disk queue disabled. all client related services are run from vm. Since reboot with all vm off does not eliminate the performance issue, only a power off did the trick, I started to think there is something up in the idrac or firmware or how p310 handles raid 1 with hotspare.
Anonymous User
4 Posts
0
January 4th, 2015 21:00
We have a server that is doing the exact same thing at a clients site. PE R420. Raid 1. Server 2012 R2 used as a terminal server. This morning about 4 minutes after a user logged in the entire server started running slow. A reboot didn't resolve the issue. I'm currently running a system state restore to a earlier date but now that i've seen this while its running i wish i could try the cold boot option before this completes....guess i'll find out shortly....
Anonymous User
4 Posts
0
January 4th, 2015 23:00
Yes i have seen this thread. The server in question also isnt running any VM's or even has the hyper-v service installed. Also the server is using a Perc H710 mini controller.
ServerDude69
37 Posts
0
January 4th, 2015 23:00
The issue may well be with using a H310 controller in this enviornment. The controller does not have write cache and therefore not designed for environments with Virtual Machines. Check your disk queue lengths to verify whether this is your issue. You may want to check out the following thread.
http://en.community.dell.com/support-forums/servers/f/906/t/19480834
I would suggest the Perc H710, which does have write cache.
ServerDude69
37 Posts
0
January 5th, 2015 15:00
The disk queue's are written before machine is shut down and therefore clear on the next reboot. Again, the H310 is not meant for an environment with Virtual Machines as it has not write disk cache. You may want to GOOGLE H310 + Performance + HYPER V before ruling this out as the issue.
cirlare
10 Posts
0
January 5th, 2015 15:00
No, that is not correct.
Think about it, if the H310 is really the issue, the symptom will be persistant.
Ever since that time it had a slow down and I hard powered it off, I have yet to re-encounter the issue. Currently i have 2 VM running on top of it over 2012 R2 standard with Hyper-V, no issues.
The best I can explain it is a firmware issue or something with disk resyncing.
Anonymous User
4 Posts
1
January 5th, 2015 17:00
Just so you are all aware, my issue has so far been fixed by going in to the BIOS (V2.2.0) and setting the CPU Performance to "Performance" instead of "Performance Per Watt (DAPC). I found that the CPU clock was limited inside of windows to 150mhz (0.15Ghz) which is why the server was running so slow. It would never increase (decrease) its throttling when the server was under heavy load. (surprising windows could run with only 150mhz). A picture of the CPU in action.
danrdj
1 Message
0
July 24th, 2015 09:00
I realize this has been stale for awhile, but it came up in my search for an answer to why my R420 VMware host's CPU was not performing well all-of-the-sudden for its two guests. An esxtop command showed very high PCPU UTIL but very low PCPU USED, meaning the CPUs were working really hard but not getting much done.
I see in April Dell released a BIOS update 2.4.2 for the R420. I'm on 2.3.3 so will plan to update ASAP.
http://www.dell.com/support/home/us/en/19/Drivers/DriversDetails?driverId=GRGC0
"Updated Intel Management Engine (ME) binary to Rev 2.1.5. 0x8B, to address issues that resulted in unexpected temporary or permanent Processor frequency degradation."
I'll also note that a warm restart didn't help, but I'm hoping a cold boot / BIOS update will.
Anonymous User
4 Posts
0
July 24th, 2015 21:00
Since I've made those changes above in my previous post i haven't had the issue re-occur.
In saying that though i'll be applying that bios update. Thanks Danrdj