Even after your MetaFrame XP server farm is "perfect," there are preventative maintenance tasks that you should perform in order to keep the farm running smoothly and to ensure that you do not find any surprises. Performing the following basic tasks will keep your system running in order:
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
- Compact the MS Access database.
- Cycle boot the MetaFrame XP servers.
- Stop and start the IMA service.
- Keep the operating system warm and clean.
- Check Resource Manager or Load Manager logs.
- Check for new hotfixes or service packs.
Compact the Access Database
If your IMA data store is Microsoft Access-based, you can use the "dsmaint compactdb" command to compact it. This should be done nightly in MS Access-based farms with more than a few servers. You can actually create a batch file and use the Windows task scheduler (or the "AT" command scheduler in NT 4) to automate this process.
In addition to the IMA data store, each MetaFrame XP server contains a local host cache copy of the data store. By default, the local host cache copy is located at the \Program Files\Citrix \Independent Management Architecture\imalhc.mdb. It is not a bad idea to compact this database as well, although it doesn't have to be done every night.
Cycle Boot MetaFrame XP Servers
Cycle booting refers to shutting down and restarting MetaFrame XP servers on some sort of predefined schedule. People usually cycle boot servers to fix any memory leaks or other random issues associated with keeping [Microsoft] servers on for long periods of time. Interestingly, Citrix feels that there is no need to cycle boot MetaFrame XP servers, but they do admit that "it might not be a bad idea" and include functionality in the CMC to manage automated server reboots.
Each time you reboot a server, the IMA service must recreate that server's local host cache which requires downloading the needed information from the IMA data store. In large environments this can take several minutes. For this reason, it is critical that you do not reboot too many servers at once. If you do, a bottleneck can occur and the IMA service might timeout on some servers.
Stop and Start the IMA Service
There are many administrators that regularly stop and start the IMA service on MetaFrame XP servers. This sometimes helps servers clear any strange problems related to IMA communication. While it is possible to stop and start this service with users connected, it's usually better done at night.
Again, you can automate the stopping and starting of the IMA service at night. Be sure that you don't have all your servers set to do it at the same time.
Stopping and staring the IMA service automatically compacts the local Resource Management database.
Keep the Operating System Warm and Clean
Because the scope of this book is MetaFrame XP, we will not address the details of tasks that should be performed at the operating system level. There are many great books on this topic. Just keep in mind that MetaFrame XP runs on top of Windows and if Windows is not cared for, MetaFrame XP will be vulnerable.
Check Resource Manager or Load Manager Logs
Checking the logs often serves two purposes. First, you can analyze the content of the logs to look for any problems that might be developing. Second, you can make sure the log files are not too large. If you find them to be extensive you can save or clear the contents.
Check for New Hotfixes or Service Packs
It is important to check Microsoft's and Citrix's website periodically for new hotfixes or service packs. For example, the IMA service that ships with MetaFrame XP experiences some trouble right out of the box. The current problems have been addressed with hotfixes and service packs and it is a good idea to keep an eye out for more.