Highlighted
Outstanding Contributor.. Outstanding Contributor..
Outstanding Contributor..
693 views

How to - Dynamic Disk Monitor

Every now and then I'd like to give something back to the posters and readers here because this forum as helped me out a lot already.

 

Credit for some posts:

http://h30499.www3.hp.com/t5/Application-Perf-Mgmt-BAC-BSM/HP-SiteScope-Dynamic-Disk-Monitor/m-p/6211741/highlight/true#M52933

http://h30499.www3.hp.com/hpeb/attachments/hpeb/itrc-875/52989/1/thread.pdf

http://h30499.www3.hp.com/t5/SiteScope-Support-Customer-Forum/Dynamic-Disk-Monitor-Best-practice/m-p/6302413/highlight/true#M2572

http://h30499.www3.hp.com/t5/Application-Perf-Mgmt-BAC-BSM/SiteScope-Disk-Space-Monitor-and-Dynamic-Disk-monitor-Percentage/m-p/6001021/highlight/true#M49606

 

Dynamic Disk Monitors

Why should you use them? (compared to like static Unix/Windows Resource monitors)

Unless you work in a very small or very structured(say Change Management par excellence) environment you will not be able to overview every single server you are monitoring - and not every change will come by your desk.

There are few worse things than finding a crashed server/application because a filesystem was full - and it wasn't being monitored at all.

 

Getting started

There is a reported issue with 11.20(see http://h30499.www3.hp.com/t5/Application-Perf-Mgmt-BAC-BSM/SiteScope-Disk-Space-Monitor-and-Dynamic-Disk-monitor-Percentage/m-p/6001021/highlight/true#M49606 ) - so I would suggest you upgrade to 11.21 or 11.22 atleast.

One of the more common issues, atleast here on the forum, was that when initially configured/tested - the thresholds would not trigger. This is most likely due to the SiteScope service not running with the right region setting.

The simple solution is to switch the regional setting to English(US) and see that the decimal/digit symbal are set right.(see http://h30499.www3.hp.com/t5/SiteScope-Support-Customer-Forum/SiteScope-11-22-Dynamic-Disk-Percent-Full-Problem/m-p/6216581#M2076 )

That should be all to get it going.

 

And off you go!

Create a Dynamic Disk Monitor, select the Server you want, hit enter. It is as easy as that generally.

If you don't need MB counters, select and delete them.

If you aren't interested in removed file systems, uncheck the checkbox to "display removed counters".

If you are fine with the default settings of 95/98 warning/error then you are good, otherwise adjust them or fix them in your template.

On a personal note: I stopped using the dynamic disk monitor in all *nix templates a while back - I manually add them and get a clear list from the guys in charge on what to exclude.

 

Regex, Regex everywhere!

Now, most posts in the recent months have been around dynamic use of regex.

/.*/percent full/

This will by default grab each and every file system it can find. Including say fixed NFS mounts, unmounted disks and everything else it can find. In many cases, and specially for some applications howevery, some filesystems are fully allocated and the application itself manages the space in the specific partition.(Oracle, SAP and liken come to mind...).

Monitoring a filesystem that always shows as warning/error has no benefit(it just costs you an extra point also...) so let's get rid of that. (Kudos again to kenneth and LGentile for this)

Windows

/^(?!.*Harddisk).*/percent full/

/^(?!.*Harddisk).*/MB free/

/^(?!.*Harddisk).*/MB total/

This will exclude all Harddisk drives.

 

Unix

/^(?!.*(/oracle/.*/oradata|/mnt)).*/percent full/

 

Say your system has

/oracle/tst/oradata1

/oracle/tst/oradata2

/oracle/tst2/oradata3

/oracle/tst1/oradata1

All 4 will be ignored since they match the pattern.

And anything thats in /mnt or possibly below will also be ignored.

 

Singularities

Sometimes, some filesystems need stricter thresholds. Stricter is easier to accomplish than more lenient.

Just hit the "Get Counters" pick out the filesystem you need and add it with stricter thresholds.(say 80/90)

The stricter threshold will always trigger over a global threshold of /.*/percent full/.

However, this doesn't apply(or atleast to my knowledge of 11.23) to more lenient(say 98/99) ones.

If you need that, you'll have to exclude the filesystem and add it on it's own - or the global rule will trigger first.

 

Words of caution

When static is good to have. I've had 2 cases where I picked a static monitor(or rather a disk specific regex) over an all dynamic regex.

One is a server with an internal backup disk that gets swapped daily/weekly.

Dynamic is great - but it doesn't tell you if the disk is in or not. If the backup disk always gets the same drive then that is rather easy - and depending on use case rather important.

 

The other case is a Microsoft Cluster. If a Resource Group has disks assigned, I specify the exact disks rather than monitoring all disks on the physical machines. This is purely by convenience I'd have to say - but it allows you to see what uses what on the cluster.(Downside, if anyone adds disks to the Resource group without saying...).

 

 

Hoping I didn't write anything wrong right now and leaving it with the hope that it helps out some more users.

 

regards

Labels (2)
0 Replies
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.