Disk Enclosures
1748140 Members
3677 Online
108758 Solutions
New Discussion юеВ

Re: va7400 Warning State

 
vladimir_1
Frequent Advisor

va7400 Warning State

The array lit orange fault led and all even disks are flashing from time to time. No other LEDs lit. The array can't be seen by CV. The only way into it is rs232 port. This way the array is seen in Watning State. All LUNs seems good (can be mounted) but I'm afraid of using them. File with some actions via serial port in attachment. Any suggestions?
8 REPLIES 8
David Uraneck
Occasional Advisor

Re: va7400 Warning State

Check for warning messages in syslog and CV status menus. The flashing on even disks indicates a problem with the controller for that redundancy group. We had something like this after a bad crash, both a controller and one failed disk were replaced.

However, we continued to have periodic synchronous flashing of all the disks amber LED's. CV reported a lot of scrubbing going on, and Warning "non redundant disk access". This turned out that C1 had a "Link Up" state on J1 port trying to link to a phantom JBOD array which did not exist. Reseating the loopback plug on J1 (next to the FC port on the back)fixed that problem with phantom link up state.
Eugeny Brychkov
Honored Contributor

Re: va7400 Warning State

Vladimir,
first of all log call with HP. Then check if you have data backup (this does NOT mean that something wrong to occur, but just good practice to be on the safe track).
Your VA has HP15 firmware, HP engineers will upgrade it to HP18, and upgrade commandView to 1.06.
From what you're saying I suspect that something is wrong at the backend of M/C2 controller. What - should be checked by HP - if this is really controller, disk or backend FC cable causing these symptoms.
You can use LUNs now. This warning is about redundancy loss (NOT data loss). It's very good that you noticed it.
What CV says when you're trying 'armdsp -i' and 'armdsp -a va1'?
Eugeny
Eugeny Brychkov
Honored Contributor

Re: va7400 Warning State

BTW, running 'recover' command WAS REALLY EXTREMELY BAD idea. VA did not have maps corrupt or lost. Running this command witout reason may corrupt data totally.
As I said - if you see something wrong with VA please DO NOT run commands you do not know. Call HP!
Eugeny
vladimir_1
Frequent Advisor

Re: va7400 Warning State

armdsp -i returns nothing. Call to HP has been logged several hours ago but no callback so far.
As long as a suspicious fru is M/C2 is it worth to pull it out for troubleshooting? About recover. As I remember, healthy array replies to recover command as 'nothing to recover' (or smth like that). Am I wrong?
And last question. What is Volume Set and its serial number?

Thanks a lot in advance.
Eugeny Brychkov
Honored Contributor

Re: va7400 Warning State

About HP responce time: if you consider your system as critical and want them to react faster then you need to upgrade support contract. Now, if you consider situation is critical, call HP and tell them that you want this problem resolved ASAP. I believe it's possible to come to agreement;
About recover: I did not try it by myself from VFP. From CV, yes, it says 'nothing to recover' or 'no need to recover'. Not sure about VFP command which may be able to force VA to reconstruct maps.
Volume set s/n is an identifier of current VA volume set. This value is written to all the disks within VA and allows any VA to which disk may be installed to recognize if this disks belongs to its volume set or not.
About 'armdsp -i': I do not know which OS you use, you need to stop host agent (HA_trigger in hpux), then stop opendial (dial_trigger in hpux), then find and move DeviceDBFile in anoter location, and then start these 2 services in reverse order. Then issue 'armdiscover' (new DeviceDBFile database file will be created). If armdiscover will discover VA, then run 'armdsp -i' and then 'armdsp -a', but if not.... then you may need HP CE onsite
Eugeny
vladimir_1
Frequent Advisor

Re: va7400 Warning State

Then why it shows different Volume Set S/Ns? Is that another result of a problem with M/C2?
Eugeny Brychkov
Honored Contributor

Re: va7400 Warning State

I do not know. It can be recovery problem. This started when you've tried 'recover -s' first time. Never run 'recover' again unless HP will ask to
Eugeny
David Uraneck
Occasional Advisor

Re: va7400 Warning State

I did not read your post closely enough-sorry for the CV access assumptions. Losing CV access would upset me--what does armdiscover show?

When we had our problem with Warning: "Non redundant disk access", it was explained to me by the support center that the warning was referring to the disks on your JBOD, not the disks in the main enclosure. The disks in the M enclosure should be fine.

Again, I think your controller MC2 is losing its connection to your JBOD for some reason. Our VA7400 does not have a JBOD attached. However, each conroller has a loopback plug in the external port which keeps M/C1.J1 = M/C2.J1 = Link Up. When C2 was generating "Non redundant disk access" errors we were seeing M/C2.J1 = Link Down. I was looking for these port states in your attachment and could not find anything.