Mailserver client failed to start/complete

[alexbplee]

1. I encountered some issues when I was tried running 8 tiles ，and there is some Warings Messages and Issues information

in my Score_x_Tile_Test.out file.please see picture “error1.jpg”and“error2.jpg”.And check the result file for detail！

2. When I try run only 4 tiles，there is no error1 and error2 ，and I can get score，but still has some Warings Messages and Issues information

in my Score_x_Tile_Test.out file.

Would you please give me some suggestions and methods to figure them out.

Thank you!

[alexbplee]

One more question，I run a 4 tiles test just now，and I got the final score ，but I got another error information

in my mailserver log file ，please see the result file！

Microsoft Exchange Load Generator

Configuring LoadGen task engine...

Loading settings...

NOTE: There are no trace listeners found. You can add/remove shared trace listeners.

Starting simulation...

Simulation has failed.

Task exceptions : 0

Simulation duration : 03:08:23

Tasks completed : 340150

Users simulated: 1000

Report file     : C:\Program Files\Exchange Load Generator\Loadgen20130816_182557.html

And still has some Warings Messages and Issues information in my Score_x_Tile_Test.out file.

Thank you！

[RebeccaG]

Hi,

Thank you for providing both the results files, they are very informative. It would be a good idea for you to set DEBUGFLAG=1 in the VMMARK2.CONFIG so we can get a bit more information.

It looks like the only problem you are really having is with your mailserver workloads.

Regarding the warnings in Score_N_Tile_Test.out, see this post:

https://communities.vmware.com/message/2238161#2238161

On the 4-tile run, everything finished normally. However, on the mailservers, you can see spikes in the Task Queue Length periodically. If the Task Queue Length reaches higher than 1500, LoadGen shut down by design. Your mailservers never reached that point, so the run completed normally. However, these spikes do generally indicate a shortage of some resource. Your Olio and DS2 workloads seem like they're doing ok; is there some resource shortage that could affect Mailserver specifically? Is it on different storage than the other workloads?

On the 8-tile run, mailserver0 ran for a short time and then quit because the Task Queue Length was too high. All of the following mailservers failed with error "ERROR -- Caught exception COMException: The server is not operational." This is why you got the message "Mailserver failed to start/complete". I think this is caused by whatever caused task queue length to spike on the 4 tile run.

You might also try deleting the "192.168.1.10 primeclient" entry from the prime client hosts file, so there is only one entry for the primeclient in its hosts file.

Thanks,
Rebecca

[alexbplee]

Thank you for your answer!

I wonder may be related to hardware resources shortage in my environment.

I have 2 physical host,and each host have 2 CPUs with 24 virtual thread cores and 128G memory,that's mean,

in my cluster ,I have a total of 96 virtual cpus and 256GB of memory to run the benchmark,each host has 4 10GB

optical fiber nic ports, 1 for management, 1 for vmotion and 2 as virtual switchs,all of those optical fiber nic ports

connect to a 10GB FC Switch.

And each host use a Qlogic 2562 8GB HBA card to connect a EMC SAN storage,there is 5 LUNs in this SAN

storage, two host share those 5 LUNs,each LUN has 2tiles.

So,any suggestion for me ?

Thank you!

[RebeccaG]

Hi Alex,

About the storage- what matters most to SAN performance, if it is HDD based, is the number of disks underlying each LUN.

1. How many disks are in each LUN?
2. We typically place one tile per LUN. How did you distribute your tiles across LUNs?

VMmark has a number of 8 tile published results: VMware VMmark I recommend looking at the hardware used in these results to get an idea of what kind of hardware is needed for an 8 tile result.

Thanks,

Rebecca

[alexbplee]

HI Rebecca,

               Thank you very much for your responsible answer.I drew a topology diagram for my environment,maybe I need reallocate my storage.

Best Regards !

[RebeccaG · Accepted solution]

Nice diagram! I see you have 4 LUNs for VMmark tiles, plus 1 LUN for DeployTemplate.

In my opinion, 9 disks per LUN, with 2 tiles per LUN is probably not fast enough storage. It would be adequate for 1 tile per LUN.

If you have not already, try this:

1. Run 5 tiles with 1 tile on each LUN (1-4) except for 1 LUN holding two tiles. (for example, tile 0 and tile 4 are both on LUN1)

2. In Score_5_Tile_Test, look at the QoS scores for the tiles which share the same LUN. Are the QoS scores higher than the other tiles? If so, this could indicate a storage bottleneck.

Also try changing Deploy/DeployLUNs="vm_lun1" to Deploy/DeployLUNs="vm_lun0". You should generally use different LUNs for DeployLUN and tile VMs so Deploy does not impact the workload VM performance.

One other thing that can cause unpredictable performance is if the server's power management setting in the BIOS is set to 'balanced' (power saving) or similar; you should always set to OS control.

Thanks,

Rebecca

[alexbplee]

Hi,Rebecca!

        Good news! I have solved this issue, because of the bottleneck of storage,thank you very much!

Best Regards!

Alex

[RebeccaG]

That's great, glad to hear you have resolved it!

Rebecca