Event ID 1065: Source= Insight agents

[jodon]

Hi

And before you all tell me this should be in the NT section, please read on.
We have a 2 Proliant 6000's set up in a cluster both attached to several raid expansion boxes via 3 (per machine) Smart 2/P array ctrs. using NT3.51. In the event viewer I have several errors with event id: 1065, source:insight agents,message reads: Drive array insight agent:IDA accelerator Status change slot 3(and 4 and 5)Value:2!=Threshold:3.
I have tried running the Compaq Array config utility in NT, but this does not show any errors and yet I suspect it is a Smart 2/P ctr. issue and not S/W.
Have checked Compaq website as well as various others without any luck. Can anybody help me pinpoint a workaround/fix, or even explain what this message means????

Thanks in anticipation
jodon

 

[gadfly]

That problem is not very well know, but happens all the time on older controllers. We ran into the same Event ID 1065 error a couple times. It turned out to be bad cache on the controller. If you run array diagnostics, it would not show up, but if you go to select cache option, and see no selection, that means your cache is bad. If your server is still under warranty, Compaq will replace it. Hope this helps.

 

[Gafly]

One more note, even if you get the card replaced by Compaq, there is a chance the replacement card will have dead batteries. They do not make the Smart 2X controllers any more, so the replacements were all refurbished parts.

 

[jodon]

Gadfly/Gafly

Thanks for the feedback.
This issue is still ongoing at present and I will use your comments the next time I have downtime to work on them.
If it would benefit others to know the results I will gladly write back under this thread if there is a response to this note. (One response would do. Don't want to create a massive build up of replies here).

Regards
Jodon.

 

[connollyg]

If you are still using this config, should you even have cache enabled?
Using Back-Plane Raid Controllers in a Cluster with cache enabled leaves you exposed to data-loss, if a server crashes before the data is flushed from the cache!

You should use external Raid controllers in cluster configs, something like the RA4100 or the MSA1000, or possibly the new SCSI cluster storage (looks like the MSA1000, but has a SCSI front-end)

Gerald