SMS server constantly hangs...unable to keep up!!!

[PhilMarcum]

I'm wondering if anyone has experienced this issue and can point me in the right direction:
After successfully upgrading my primary site server from SP1 to SP2 I notice that it hangs for an unusuually long time. It gets so that once the server hangs no one in the entire organization is able to log in to any of their consoles. Checking the event logs I show event ID 2021 & 2022. I followed the instructions and modified the registry as outlined here:

http://support.microsoft.com/?kbid=317249

Rebooted the server and the problem happened a few days later and is still happening. My hardware:
Dell PE 4600 w/4GB RAM etc...
I have over 5000 users and close to 6000 systems. My SMS database is up to 30GB and all of this was installed prior to my arrival.
I'm pushing for another server but management isn't biting right now. So would anyone know of anything else that can be done to help figure out this issue?
Any responses are appreciated.

 

[ncotton]

Have you checked your system stats when this problem is occuring, any overhead processes, memory usage, cpu usage?

Hope this Helps.

Neil J Cotton
njc Information Systems
Systems Consultant

 

[PhilMarcum]

Unfortunately not because use the SMS admin console so I'm hardly ever logged directly into the SMS site server. I'll get a call from the helpdesk or a fieldtech notifying me of it if I don't catch it myself. The other problem is that when the server starts to hang it's hung. Nothing responds it just freezes up and I have to either hard boot it or wait for it to clear itself up.

Thanks

 

[ncotton]

You don't have any ridiculously large advertisements scheduled at the minute do you?

Hope this Helps.

Neil J Cotton
njc Information Systems
Systems Consultant

 

[PhilMarcum]

No large advertisements or anything new since upgrading to SP2.

THanks

 

[tbrennans]

Thats a stupid huge db for that many clients:

Check that the DB is getting backed up, its probably all transaction log.

You probably have your DB recovery set to FULL on the sql server

change it to a simple recovery model, this in effect enables " Truncate log on checkpoint" also check the box for " auto shrink" ...to reclaim that free space.

(this is done within sql enterprise manager..right click on the sms db and choose properties).

 

[tbrennans]

that might be part the other might be dupes bad ddr whatever:

check out this to find out what is causeing your proc to max out on you:

http://support.microsoft.com/?scid=kb;en-us;837372&spid=2889&sid=global

I'd also check the different inboxes and see if one is clogged up, if so move all that to another folder after stopping the sms services. and see if it takes off from there

 

[PhilMarcum]

Turns out I'm able to get another server, Proliant DL585, out of this and now need to proceed with steps to decommission this Dell 4650 and get SMS up and running on this new server before Monday. In preparing for this I see that the reason for the large database is pointing to the \inboxes\auth\dataldr.box. This folder is now showing as 26.4GB and contains nothing but .MIF files. What I need to do is gather a decent backup or snapshot of this site server before taking it offline and then use this to restore to the new server. I'm searching but can't see a need to try and backup all of this up. Can someone confirm this for me.
Also if anyone have any advice or steps on restoring SMS to new hardware I'd be greatly appreciative.
Either way it'll be a long weekend for me and I'm sure most are preparing for theirs. Therefore ANY responses are appreciated.

 

[tbrennans]

well, in this situation I'm fairly sure MS would tell you to copy the mif files into another folder. then delete all of them OR just delete the 1st few that are gumming up the works, then you should be good again.

Take a look at this 1st..

http://support.microsoft.com/default.aspx?scid=kb;en-us;201544

 

[tbrennans]

when I say "then delete them" I mean the ones in the inbox, not the ones ya just copied into some other folder :-D

 

[PhilMarcum]

If I copy the mif files into another folder and delete the original ones can I then back up the SMS site server, re-install my new SMS server, restore Sitectrl file, SMS & NAL registry keys, and then copy these mif files back over to this folder, a few at a time? Is there some process for handling these files when using different SMS platforms? Or am I fine as long as I'm using the same OS ver and SP level along with the same SMS site server name and SP level?

 

[tbrennans]

no what I was getting at was there is a problem with one or some of your mif files, so if you delete the "oldest" files at the top of the folder it will probably solve your problem. delete you badmif files also. If deleting some at the top didnt work then think about tossing them all.

With that many files in there its safe to say you could start by tossing the 1st 100, then half of them, as I'm sure they have to be in there from the same machines many times over.

If you are going to have a new server, yes I would just go about a hardware replacement, as it seems you have good grasp on. I would not replace those files into the new server though. Just let it start over. (thats just me though, I tend to like to start with a clean slate)

 

[tbrennans]

One last thing, you need to know why its doing this, have you checked for duplicate guids?? this would be a reason this is happening in the 1st place.

Tim B

 

[PhilMarcum]

What would be a good tip for installing a new instance of SMS 2003 SP2 on the new server while keeping the old server online? I've renamed the old server and I'm using the same name on the new server. I've also stopped and disabled all SMS services on the old server, yet when trying to run the install on the new server it detects the existing primary site and only gives an option to install a secondary. Even though I'm using the same server name do I still need to remove the current attributes from Active Directory before I'm able to install the new primary site server? Or do I need to uninstall SMS from the old site server before installing it on the new server? I've already tried to run the exadsch command from the new server and confirmed in the log that the schema modifications were successful. Successful meaning it detected the existing attributes and reported successful.

Thanks

 

[tbrennans]

If you go into the recovery expert it will address this issue for you.

I think the sms site name is causing the problem, not the bios name.

 

[PhilMarcum]

Turns out the CD I was using to install SMS with wasn't the OEM CD and instead was SMS SP2. Once I obtained the original CD I was able to perform and complete the install.
Many thanks for all of the help and responses.

 

[mib147]

I had this problem with a Windows 2003 server. It was running SMS 2003 SP1. I didn't upgrade it to SP2 but it just started happening out of the blue. We upgraded the server OS to SP1 and it went away. Has been find ever since. Not sure what OS you are running.