storage timeout problem - esxi 5.1 fusion-io hp proliant dl380p g8

[datasecurity]

Hello!

Setup:

     HP Proliant DL380p g8

     HP 785GB Multi Level Cell G2 PCIe ioDrive2 for ProLiant Servers

     ESXi 5.1.0 799733

     HP IO Accelerator for ProLiant Install Package for VMware ESXi 5.0

The System was built around OCT'12

server performing stable and very fast - never any purples

but today, last friday and about 1.5 months ago the VMs did not repsond.

connecting with vsphere client. is ok - internal SAS storage is also ok - VMs using internal SAS are storage up and running

VMs using the fusion IO are halted - unable to access the disk.

browsing the storage shows the fusion io to be "empty" - no error only the root folder existing.

power down the machine, reboot successfull - fusion io storage is back - directory containing all folders, VMs up and runnning

the log displays:

"info : due connectivity issues no access to volume < .... > (datastore ssd) possible. trying to reconnect. result soon"

(sorry for chunky translation - the installation language is not english)

there are no newer drivers für the fusion IO, the patchnotes from esx updates dont show any storage related stuff, firmware should be also up to date.

anyhow - im going to update this weekend.

has anyone any ideas / help for my problem

[Schorschi]

What do you see for status on the datastore?  Under storage view?  What do you see for Storage Adapter status?  Have you installed ioSphere VM and the vCenter Plugin for ioSphere?  You can get explicit status of the Fusion-IO device from the ioSphere VM, Fusion-IO provides the VM as canned appliance that can show in great detail what the status of the Fusion-IO device is.  However, that same ioSphere CIM/SMI-S provider does not integrate well, I have found with enterprise management frame works.  The VM appliance communicates status via electronic mail?  No SNMP or CIM event forwarding, so an enterprise monitoring solution like Tivoli or CAUnicenter may not be an easy integration effort.  But in your case, you should be able to interactively see if the Fusion-IO device is aging or showing memory errors.  Solid State memory has a finite life, or the controller on the PCIe side maybe starting to fail for some reason.  We have been testing and plan to do more testing with Fusion-IO devices under 5.0 Update 2 and later, the big issue we have found so far, is the monitoring gap.  That is a big issue for us.