Quickplace Basic OS (Windows Focused)

Before I look into an application issue I like to start by ensuring the basic OS is there. There are many stats on a domino server that tell you how the OS is performing. It could be Domino giving issues or someting else like a monitor or backup service. What I do is to put the commands into a domino java agent so that I can build a picture in a database of the OS measure and alarm on any issues I see.

For example I would go onto a server and issue the command:-
Free Disk Space
Disk Queue Length



Free Disk Space

Show stat disk.*

Free disk space jpg







This will show you how much free space is on the disk drive. Quickplace can soon grow and use up a lot of disk space and you need a good 20% free for defrags and to give you warning when you need to bring new servers on line.





Now automate this task:-

You know the status of your disk space this next bit will help you keep on top of it going forward. The point is you don't want to log on every day and issue this command as you will forget and there is this new invention called programming that will do all this for you on a regular basis.

Daily report


I have an agent that goes out once a day and asks all servers "show stat disk.*". When the Java agent gets the result it has a pre-defined % to alarm on. In this example its 10% and you can see the red cross shows that a disk on a server has gone below the 10% number. We now just need to visit the database every so often and/or can have the database email us when a disk goes red.

Daily report





This is the form from the view:-

When the Java agent gets the result it has a pre-defined % to alarm on. In this example its 10% and you can see the Disk E is 9% so its entry is in RED. This is good. We now just need to visit the database every so often.















Now we want to know the time line of the change in disk space. If we have a reported disk that is less than 10% we create with Java a different form that will put the data this next view. This gives you the server and then the disk on that server and the free %. It lets you see the rate of decline so you know how urgent the need for new free space is.
With all this you now are in control of your disk space and the days of finding out you have 3 Gig left on a drive should be over.

Disk Splace List

Disk Queue Length

Show stat platform.logicaldisk.6.*

Disk Queue Length You need to work out what are the disks you are interested in. In short you want the data disk. In this example its disk E. You can issue a "show stat platform.logicaldisk.*" to find all your disks and then look at Platform.LogicalDisk.6.AssignedName" to find the disk number you are looking for. In our case our disk number 6 is our E data drive. You don't want anything else on this disk. i.e. OS or log files or Transactional logs. You want this to be a disk on its own spindle. Disk Queue Length shows you how busy the disk is. If you look at AvgQueueLen = 0 this shows at the snap shot nothing is waiting to access the disk. Also the pctUtil is 1.3 so that only 1.3% of the disk time is being used. But don't let a snap shot fool you. The real numbers we are interested in are the AvgQueueLen.Avg . This is over a period of time and is a better indication. You want an AvgQueueLen.Avg of 1.2 max per spindle on the drive. Yes only 1.2 max. I have seen systems with a value of 40-100 on this. In this case quickplace was just ripping itself apart. Compacts were talking 7 days to complete and a newsletter 4 days. Thus a news letter was kicking off before the previous one had finished. We will come back to this value later but on a windows system this is a real indicator to the work load your disk are going through and rest assured Quickplace will kill your disk if you let it.