Video Screencast Help
Endpoint Management Community Blog
Showing posts tagged with Performance
Showing posts in English
riva11 | 06 Mar 2008 | 0 comments

Nel caso sia stato creato un ambiente di pre-boot Linux (PXE, Automation BootDisk, Network BootDisk, ecc), ma per qualche problema non viene montato un disco che punta ad una condizione sul server DS. Come è possibile collegare manualmente da prompt"/#" una condivisione in modo da fare un test?

Eseguire il comando seguente:

"mount -t cifs -o username=[nome utente],workgroup=[workgroup/dominio],password=[password] //[dsname/IP/FQDN]/express /mnt/ds"

Se il comando ha avuto successo, viene visualizzato un prompt "/#", altrimenti si ottiene un avviso su schermo o un messaggio di errore.

Per controllare che sia stato montato correttamente, digitare "mount" che visualizzerà la lista di tutte le cartelle, device, ecc.

"cd mnt/ds" fornisce inoltre l'accesso alla cartella condivisa del server DS.

Digitando successivamente il comando "ls" sarà fornita la lista delle directory della condivisione sul...

SK | 05 Mar 2008 | 5 comments

A Linux pre-boot environment has been created (PXE, Automation BootDisk, Network BootDisk, etc), however, when it runs it fails to mount to the DS. How can the mount process be tested from the "/#" prompt?

Run the following command:

"mount -t cifs -o username=[username],workgroup=[workgroup/domain],password=[password] //[dsname/IP/FQDN]/express /mnt/ds"

If this action is successful, the "/#" prompt will return, otherwise an error or usage screen will be returned.

To check the mount, type "mount" which will display all mounted devices, directories, etc.

"cd mnt/ds" will provide access to the DS mount point.

Typing "ls" will provide a directory listing of the DS mount point. | 05 Mar 2008 | 2 comments

You just installed a Hot Fix that included a new AClient and you notice that as Auto Update is enabled, it only updates 15 clients at a time. You may find yourself asking, "How can I change the AClient Auto Update value from 15 clients at a time to a different value?"

When a newer version of AClient is installed into the Deployment Share, the managed computers automatically update their copy of AClient to match this newer version. Starting with Deployment Solution 6.5 Hotfix 2, the AClient Auto Update feature only updates 15 computers at a time, to reduce the network traffic used when a new version of AClient is installed.

To view or change the auto update value:
On the computer that Deployment Server is installed on, run Regedit.

Locate and select the following registry key:

HKEY_LOCAL_MACHINE\SOFTWARE\Altiris\Altiris eXpress\Options

Locate and double-click the UpdateAgentsConcurrentlyMax DWORD entry...

CondorMan | 29 Feb 2008 | 5 comments

If you plan on installing Vista Service Pack 1 when it is released (highly recommended), you might want to run the cleanup command afterward.

To do this, open the run dialog (Win-R) and type "vsp1cln.exe" (without quotes) and press enter.

This will remove the ability to uninstall the service pack (so make sure to test everything before running the cleanup), and it will cleanup 1GB worth of data off of your hard drive.

BRING | 21 Feb 2008 | 0 comments

Measure twice and cut once. Sometimes not remembering to do that causes trouble, and has sometimes ruined some beautiful wood stock.

Frequently we do something similar in the Task Server environment when we forget to check the collection references that our tasks are assigned to. Like sending a Shutdown, or Restart task to "All Computers".

While this won't be painless to all end users, we can mitigate it and stop it from hitting everyone, using the following procedure:

  1. 1. Delete the Task from the Console. This may also involve deleting any referenced items as well.
  2. 2. Restarting the Altiris Object Host Service. This will clear the Task Server cache, and allow any deleted jobs to be cleaned and not promulgated to other machines.

It's just that simple. And while it won't be painless for you as the...

BRING | 13 Dec 2007 | 0 comments

Why are my SQL server transaction logs growing each day to the point they overrun available disk space? This is really getting annoying. I am having to truncate them manually. Why isn't the automated procedure working?

After checking EvtQFast and EvtQueue thread settings, and adjusting the Throttle delays as found in KB 2344 and KB 17079, the problem just won't go away. The transaction logs were getting still getting a lot of pressure.

We are going to need some more information. So we set the following Performance Monitor counters on the SQL Server:

  • SQLServer:Buffer Manager: Buffer Cache Hit Ratio
  • SQLServer:Buffer Manager: Checkpoint pages/sec
  • SQLServer:Buffer Manager: Page writes/sec
  • SQLServer:Databases: Log Bytes Flushed/sec (Select correct Altiris database...
BRING | 29 Oct 2007 | 1 comment

We all know that there are big Software Delivery packages in the world. We also know that sometimes network infrastructure lets us down. So what do you do when you have this monster of a software delivery package to get to clients, and the network hiccups in the middle. Well ... it depends!


Is there anyway to get Software Delivery to resume a package download from an intermediate state, rather than restarting the package download over completely, when the package delivery mechanism is UNC? What about HTTP?


If you are using UNC, then the bad news is NO. This is called package checksumming and Checksums are only available via HTTP downloads, not UNC. There are two other things to note here.

  1. Once a solid network connection is restored,if your package contains multiple files, the download will resume at the end of the last file successfully delivered, but only via HTTP. This is not supported on UNC deliveries.
  2. ...
Andrew Souter | 27 Sep 2007 | 3 comments

There is a known problem that sometimes arises when using Summary report definition. The COUNT does not report correctly but when drilling down the next level of the report is correct. For example the summary level of the report might show 58 computers of a particular operating system but when you drill down there are 65 computers listed, which is the correct number.

The issue usually surfaces when selecting data from multiple tables. To limit the data to the resource type selected, a view called vResourceEx is used. When fields are selected from the other tables the combining of the data from the view and the other tables is not done properly. It is simple to correct the problem with just a few steps using the Advanced edit button which opens the Report Query Builder.

  1. Open a summary report
  2. Click Edit this report....
BRING | 26 Sep 2007 | 0 comments

A very pesky thing in Notification Server is schedule that Purging Maintenance runs on. The default value is that it runs every day at 2:30am. So you have a good sleep going, and all of the sudden, it is rudely interrupted by someone calling you and telling you that your Notification Server is deadlocking all over the place!

Not good!

There are so many other scheduled events running around this time, and of course, Purging is a very database intensive activity, and would well be left to a timeframe when no one else is accessing the system.

Well, while we can't control everything, one frustration has been being able to change the Purging Maintenance schedule. Some have tried to change it via editing the scheduled task, but to no avail.

While a clean, user-interface method is still in the works, a quick and fast solution...

SK | 20 Aug 2007 | 0 comments

Patch Management Solution sometimes encounters problems when additional updates are created by Microsoft or when revised versions are released or even when the NS fails to download all of the updates the first time.

These scenarios will prevent you from creating a SW Update Task for the Bulletin until the missing updates have been downloaded.

I have quite frequently experienced that when using the "Recreate Packages" option, the NS still fails to download one or two of the updates.

When this happens, I simply disable the Bulletin, and then stage it again, as this always downloads all updates successfully.