one node shuts down

[nipun_2]

Hi,
Since two days this problem has started.

We have a common environment cluster with (V7.3-2)
Server (DS25)
and the following nodes
Sat1- also contains data disks) (DS25)
Sat2 - satellite node (DS25)
Sat3 - satellite node (DS25)

I need to bring some data back from tape and put it on Sat1. So every night I start the tape backup and when I come in the morning the node is shutdown. I have to press the power button and start the node.

Could this be a hardware issue? or plain software issue. We recently have upgraded to V 7.3-2.

However,after the upgrade when I bring data back from tape or putting data to tape seemed to work fine before.

Any suggestions or comments please let me know
Nipun

[Karl Rohwedder]

Have you checked the OPERATOR.LOG and the ERRLOG.SYS? Any strange entries?

regards kalle

[Stuart Fuller_2]

By "shutdown", do you mean that the system has returned to the console prompt (>>>), or do you mean that the power is now off. I note that you said that you have to press the power button.

If the power is off, then you may have a bad fan (causing the system to overheat) or a bad power supply. In either case, it's most likely a hardware problem.

[nipun_2]

Hi Karl and Stuart,
Thanks for the reply.

I went to the SAt1 node and tried to search for
sys$manager:operator.log but there was no log file.

There was a operator.log file on the server node however it seems to have all enteries specific to itself.


I have not checked errorlog.sys yet.

Stuart, the system was completely off so no It was not at the SRM Console (>>>_) promtp.

I had to press the power switch to turn it on. I also think it could be a power supply problem. There is a way to further analyze this please let me know.

I am still a bit new to openVMS so if you guys could give me some detailed commands to analyze and view/post certain file it would be very helpful.

[Lawrence Czlapinski]

Nipun: Since the power button is off, it sounds like a hardware issue. If it was a software shutdown, it has been my experience that the system normally would wind up at the >>> prompt or it would reboot itself.
1. CHECK THE POWER CONNECTIONS: It's unlikely but jingle the power cord end at the computer. On June 9 th, I found a found a power receptacle on a VAX 4105A had too much tolerance. I could actually jingle the power cord in it and the power would go off or on when I moved the power cord in it. The computer had been dropping out of the cluster at random intervals.
2. Check the CPU FAN(S). The CPUs could be overheating.
3. Does jingling the top of the CPU cabinet affect the system's front panel? We once had a bent INTERLOCK pin. When it lost contact the system died.
4. It could be that the CPU POWER SUPPLY overheats and shuts down the system.
5. It could be that the VOLTAGES at the CPU box drops too low and the system shuts down.
Lawrence

[Jim Geier_1 · Accepted solution]

We had a DS20e that would occasionally power itself off. No warnings in the operator.log, no errors in the error log file. Almost all of these events occurred when backups were running to an SCSI-attached TL892 or later to a MSL5026.

After lots of attempts to fix the problem by replacing the SCSI adapter and power supplies, the fix was to replace the front panel switch. Once that was done in March, the system has not been down once.

Our service personnel got the idea for the replacement of the front panel switch from a similar problem on a DS25. I recommend asking your service people to replace the front panel switch on the DS25.

[Ian Miller.]

SHOW POWER on the console at the >>> prompt (also accessable via WBEM) show show fan or PSU problems.

[nipun_2]

Somehow the problem has not re-occured so for now I am waiting for it to reoccur.