0

I have an HP Proliant DG380 G7 dropping one of it's arrays every hour, on the hour, for 2-5 minutes.

The OS is SBS 2011 Standard, the servers runs Exchange, DC, files & Trend WFBS 8.

I can watch the D Drive disappear for the duration of the problem - then it just comes back up and all is well again. There is no loss of network connectivity, although the mapped drives also disappear.

We thought it might be to do with Sharepoint / VSS writers failing but it looks as though this is a symptom rather than cause. It survives a reboot.

Any ideas as to what could be running on a regular schedule like this?

ewwhite
  • 197,159
  • 92
  • 443
  • 809
jidl
  • 151
  • 1
  • 2
  • 9

1 Answers1

0

Battery and flash-backed write caches (BBWC/FBWC) are very important when working with HP Smart Array controllers. Performance tends to suffer without them. I suspect you may have a few things happening here...

RAID5 is typically a no-no. Without any write cache, you're susceptible to I/O wait and disk rotational latencies during writes. Whatever is happening hourly is triggered by your software/OS, as there are no scheduled RAID controller tasks that you'd be able to see. If the RAID5 array is unhealthy or experiencing a failing (but not failed) disk, that's even more trouble.

You can temporarily disable the controller write cache barriers to see if that's the issue. In the ACU, you can set the write cache to work even if the BBWC or FBWC is missing. It's dangerous but you can run it long enough to see if the drop-outs persist.

ewwhite
  • 197,159
  • 92
  • 443
  • 809