ESX 4.0 U1 with EVA8400 VMs Intermittent Hung

[Sivaramakrishna_1]

Hi,

We are facing some issues with ESX4.0 U1 which is connected to EVA8400. Servers are BL460c G6 with Latest firmwares.

around 25 Virtual machines across 5 ESX SErver Cluster.

Intermittently we are facing issues with VMs like a hung state for some seconds and resuming.

at the particulatime when observing the performance graph both DISK and Network reaching 100% and coming back to normal.

EVA XCS is 9534 and set the LUN policy to Round-Robin.

Can anybody help on this? Any known issues with ESX4.0 U1 etc.

Thanks,
Siva

[Antonio Milanese]

Hello Siva,

yes there are some problem with RR and ESX4.0 U1..check this post:

http://ict-freak.nl/2010/02/25/vsphere-apd-bug-is-solved-in-patch-esx400-200912401-bg/

my suggestion is to upgrade to U2 ASAP because it's more stable/predictable in EVA environments (4.1 it's even better but there are plenty of other problems with FLEX-10,ecc)

AH..don't forget to follow HP best practices when setting RR (i.e. IOPS=1,ecc) or you'll lot of ILT trashing your performance expecially on CA setups!

Best regars,

Antonio

[Jan Soska]

Hello,
we have very similar setup but with ESX4u2, no problem. Go for update 2.

To Antonio - we consider moving to ESX4.1 on new blades (including flex10 VC). You mentionned some problems related to flex10 - could you provide more links/details?

Jan

[Sivaramakrishna_1]

Hi,

Thanks for your replies. After a long time by doing serial analysis of hosts one by one, we identified one Host HBA is causing the issue and issue resolved after replacing that FC Cable.

VMware suggested to go for ESX4.1, need your additional help about the issue on ESX4.1 as this environment is running on Flex-10 VC.

Thanks,
Siva

[Antonio Milanese]

Hello,

@Siva,@Jan

yes unfortunally (HP/VMWARE..yes i'm a polite dude=) there are some major problems with flex nics drivers(bnx2x) that render the whole FLEX-10+ESX v4.1 setup unusable..

YES a REAL showstopper!

i've created a thread sometime ago and perhaps we're near to the solution, FINALLY:

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1437616

In the near future i'll stick with U2 with my customers, at least until the v1.60 driver it's THROUGHLY tested/validated!

Best regards,

Antonio

[Sivaramakrishna_1]

Finally we got the cause.. One cable which is connected from one of ESX to SAN Switch caused the issue on EVA which is resulting the issue..

Replaced that cable and everything is fine..

But still some error logs of some sense key logs appearing on vmkernal logs, which is suggesting to upgrade to 4.0 U2 to get rid out from that errors.

link is fyi
http://communities.vmware.com/message/1484442

Thanks all