Disk Arrays
cancel
Showing results for 
Search instead for 
Did you mean: 

where is the auto raid 12H metric data?

SOLVED
Go to solution
VINCENT SPURGEON
Frequent Advisor

where is the auto raid 12H metric data?

Whenever I issue...

# arraydsp -r 11270800 11271500 $ID
Performance log data is unavailable for the interval specified
arraydsp: Error in command execution

This raid array has been up and running for several months. I can't find an option for ARMserver for generating the metric data.

Where is the metric data stored? How is it gathered? Are there settings for retention of data? Etc...

Thanks!


It's only a flesh wound...
5 REPLIES
Steffi Jones_1
Esteemed Contributor
Solution

Re: where is the auto raid 12H metric data?

The arraydsp -r and -m commands rely on "clean" data in the logs to run. If there are any invalid entries in the logs, the error will be received when running arraydsp:



The following error may also appear in the /var/adm/syslog/syslog.log
file:

Nov 2 14:50:27 mims2 syslog: HP Advanced Disk Array serial number ?:
Server software logging suspended


Several events can cause the "Invalid" entries:

1. An system initialization anytime during the time period requested.

2. No data transfer during the time period requested. The performance algorithm requires at least two hours of data and looks at every
hour of data after that to compare and report performance. At least one I/O has to be logged each hour of the time period requested.

3. The ARMServer process was stopped or not logging during the time period requested.

4. Specifying invalid times when running arraydsp.

In some instances, running logprint to verify the entries in the log file will give you corrupted output (verifying that either the catalog file (LOGCATLG) or the log files is/are corrupted) - which is the reason
why arraydsp cannot list the entries for the time period specified.
This will require removal of the log files and catalog file
(see steps 1-6 below).

However, even if the LOGCATLG file appears valid, and entries in the log files are viewable, these error messages can still be received.

Perform the following steps to resolve the errors:

1. Stop ARMServer process:

/sbin/init.d/hparray stop

2. Make note of the /var/opt/hparray/log/LOGCATLG
permissions/ownership.

3. Execute the "> LOGCATLG" command to clear the file contents.

4. Restart the ARMServer process:

/sbin/init.d/hparray start

Run performance metrics when more history is available in the log files.
If there are only a few logs with invalid data, those logs should eventually
fall out of the requested time period, and valid performance data will be
available.

Andreas Voss
Honored Contributor

Re: where is the auto raid 12H metric data?

Hi,

Logs are generated from ARMServer and stored under /var/opt/hparray/log.
With logprint you can look what logs were written.
See 'man ARMServer' and 'man logprint'

Regards
VINCENT SPURGEON
Frequent Advisor

Re: where is the auto raid 12H metric data?

Found the files and tested the above on another server. Works great! Thanks!

Another question, documentation regarding ARMServer and arraymond state that if it is not running that it will NOT effect I/O on a system using the array, but will executing the "/sbin/init.d/hparray stop|start" effect I/O.

The system I really need to know the metrics on is a 24x7 system
It's only a flesh wound...
Dan Rosen
Frequent Advisor

Re: where is the auto raid 12H metric data?

Would arraylog -C $id work as well? Or are these different logs that this command addresses?
Patrick Wallek
Honored Contributor

Re: where is the auto raid 12H metric data?

Doing an /sbin/init.d/hparray stop will not effect the availability of the disk array. You will still be able to access the data.