need some troubleshooting help on my esxi box...

H

haunter

[H]ard|Gawd
Joined
Jul 20, 2011
Messages
1,883
I will post some logs later, when I can dig them up.......


Here is the issue.

I was getting a lot of logs for terrible disk latency accessing datastores, like in the 20k ms range

Eventually I dropped my VM datastore completely back in Oct

(yeah I use the VM's a lot...clearly.....but my media server is on it and I started needing it again :D )


Yesterday I swapped out the perc 5i for another perc5i and everything came back seemingly happy

minus the VM freezing randomly for a few seconds.

I figure I need to make them do some work and then check the logs again and see if I get some errors......

running 5.1(yeah I need to patch up...bad...)

its a amd 880 chipset, with a thuban x6 at 2.2ghz and 8gb of ram

the vm is server 2008r2

disk setup is...
esxi on a flash drive
vm datastore is 2x500gb sata in raid 1
my storage datastore is 4x1.5tb greens in raid 10, which are mounted to the server 2008r2 vm

I have a Linux vm and a server 2012 vm but I shut them off to aid in troubleshooting

I have literally never had to troubleshoot anything in esxi before.....ie I'm sure I've had issues but never bad enough I noticed....

posts to RTFM will be noted and are deserving...
 
Last edited:
Next time there's a freeze, go look at /var/log/vmkernel on the host - that'll tell you what the hypervisor says.
 
It may also be the Greens stalling out - those are energy saving drives, and sometimes act really "strange" on RAID groups, as RAID doesn't expect a drive to idle down.
 
yeah it could be.

This setup is about 4 years old and the issue is a few months old at this point. They do wake up slow at times, and I fully expect that since they like to park and idle down, so I'm hesitant to blame to the drives for it, *yet*

I hope to replace them in the next year with something a little more suited to what I need.

I had 2 of them from just needing a JBOD for storage, so it made some sense to just get 2 more for this, way back when, half the raid 10 reasoning was to help make up for the slowness of the drives
 
the Perc 5i is using writeback with the BBU (Battery Backup Unit) correct?

Make sure the battery is reporting good and if you dont have one i have seen plenty of issues where without the BBWC the latency is hell....

How is your raid configured in the Perc (write-through/back?)
 
dasaint said:
the Perc 5i is using writeback with the BBU (Battery Backup Unit) correct?

Make sure the battery is reporting good and if you dont have one i have seen plenty of issues where without the BBWC the latency is hell....

How is your raid configured in the Perc (write-through/back?)
Click to expand...



yeah I have the battery installed.

I will have to double check the config, I believe I just made sure it was factory defaulted and then created my disk groups etc
 
I had one time where my controllers defaulted to Write-Through instead of Writeback, also make sure the firmware is good and your controller isnt running a batt test... seen all of these put the controller into a duller mode.

I use the LSIProvider VIB on my hosts then i can use the LSI MegaRaid Storage Manager to Control/Monitor my Dell Perc H700 Controllers... dont know if the 5's would work but if they are LSI based it might work :)
 
2014-12-08T22:41:08.098Z cpu1:2049)NMP: nmp_ThrottleLogForDevice:2319: Cmd 0x85 (0x41240080c9c0, 3170) to dev "naa.600188b04cd96b0016dd59139186bf59" on path "vmhba2:C2:T0:L0" Failed: H:0x0 D:0x2 P:0x0 Valid sense data: 0x5 0x20 0x0. Act:NONE

2014-12-08T22:41:08.098Z cpu1:2049)ScsiDeviceIO: 2316: Cmd(0x41240080c9c0) 0x85, CmdSN 0x222 from world 3170 to dev "naa.600188b04cd96b0016dd59139186bf59" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0x5 0x20 0x0.

2014-12-08T22:41:08.098Z cpu1:2049)ScsiDeviceIO: 2316: Cmd(0x41240080c9c0) 0x4d, CmdSN 0x223 from world 3170 to dev "naa.600188b04cd96b0016dd59139186bf59" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0x5 0x20 0x0.

2014-12-08T22:41:08.098Z cpu1:2049)ScsiDeviceIO: 2316: Cmd(0x41240080c9c0) 0x1a, CmdSN 0x224 from world 3170 to dev "naa.600188b04cd96b0016dd59139186bf59" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0x5 0x24 0x0.
 
perc 5i package 5.1.1-0040
fw ver 1/03.10-0216
bios cer mt28
ctrlr version 1.04-017a
boot block ver r.2.3.12
 
I got the fw updated, the vibs installed.

I just need to work out getting MSM to see the card, via my server VM, so I can have some more visibility
 
the cims server settings never show up anywhere that I can find

even tho I installed the vibs, might need to try a different version of MSM, tho
 
Also have seen issue where you need to disable full firewall to get it working.. Try below first then if it still not showing then try disabling FW

in the MSM you click Configure Host -> Display all ESXi-CIMON servers in the network of the local server -> Save Setting -> Click Discover Servers

You should see your hosts then...
 
thanks!

I didn't get to look at it last night, had maintenance all night. maybe tonight
 
firewall disabled,....MSM still isn't seeing it

even newer firmware installed

trying some older MSM install....
 
You must log in or register to reply here.
Back
Top