AD falling over; DNS falling over; DHCP falling over!

[downloadkid]

Afternoon everyone, got a strange problem I can't seem to resolve, so if any of you gurus have got any input that would be greatly appreciated!

the problem;
2 weeks ago for no apparent reason all clients lost connectivity to the DC. when i looged into the DC and bought up the AD users and computers snapin it said' server is not available'. when i pulled up the DNS and DHCP snapins that message was there was insufficient 'storage' available. Same result on the second DC.
the event viewer gave no clues as to the problem, but listed a series of errors, like netlogon, unable to apply GPOs as server not available, file replication error etc.
rebooting the server bought it back up, but i also had to reboot all the other servers to 'reconnect'. this goes to for the clients.
in the client logs the message is the same, namely insufficient storage!
there is plenty of storage on both the DCs so I though it might have been page file, which I increased which I thought had resloved the probel m until today......
In the middle of ghosting around 30 pcs, ghost slowed to a crawl to the point where I had to stop it after which the server issue returned.
Has anyone experienced any thing similar or have any 'polite' suggestions?

Many thnaks in advance.....

 

[downloadkid]

......sorry forgot to mention;

I've run DCdiag and Netdiag, there are no errors.

 

[Lemon13]

just a guess, check the registry size / limit on dc

M. Knorr

MCSE, MCTS, MCSA, CCNA

 

[58sniper]

what else is running on the DCs? Are both GCs?

Pat Richard MVP
Plan for performance, and capacity takes care of itself. Plan for capacity, and suffer poor performance.

http://www.ucblogs.net/blogs/exchange/

 

[downloadkid]

Afternoon people, thanks for the input;
i haven't checked the size of the reg..........I'll look.in to that.

Only one server is the GC, the second is also an application server - we're a school.

Today was a nightmare with AD falling over around 3 times, each time having to reboot the servers to bring the network back up. Both servers are running 2003 R2 x64 with 4 GB of RAM, plenty of storage space etc. I edited the reg adding the IRPStackSize or 18 in the hope that that would resolve it - it didn't.
when it fell over I ran DCdiag - it errored saying the DC was not available. I then ran netdiag which errored saying';
error_notenoughmemory!

the only app added tothe server in the last month was Etrust ITM agents (AV) which I have removed from both servers.

It hasn't fallen over since, but then by the time it was removed the servers weren't under much load.

I shut all the servers completely down then systematically bought them back up - all seemed to be well until event viewer on the GC DC showed errors 1058 and 1030 - basically saying that the Dc was unable to access the gpt.ini in the sysvol, access denied.
Yet if I type in the FQDN and path to the gpt file on the DC it can access it without error. I checked the permissions on both the sysvol and the root of the drive and are correct.
So I rebboted the GC DC again and the error hasn't returned.
I performed a GP logging on the DC with no errors.

I won't know until it's under load tomorrow if the problem is still lurking.

Incidently the GP errors were not duplicated on the second DC.

 

[58sniper]

You should make all DCs into GCs.

Pat Richard MVP
Plan for performance, and capacity takes care of itself. Plan for capacity, and suffer poor performance.

http://www.ucblogs.net/blogs/exchange/

 

[downloadkid]

well today seemed to be OK; removing the AV sorted the issue.
I ran Poolmom fro most of the day to monitor paged and non-paged memory.
I have to say i was shocked at how much memory gets 'leaked' but all manor of processes............you live and learn!

 

[downloadkid]

SOLVED>>>>>>>
Truned out the some managment software from Symantec / Altris had decided for some unknown reason to hog ALL availabe user ports on each of the servers the Altris Agent was installed on. I have no idea why, I've been using it for a few months without issue; removing the agent and rebooting set everything back to 'normal', a netstat -b showed that Altris was no longer blocking communications.
So when you get a error along the lines of ;insufficient resources to run command' or'not enough storage is available', or 'insufficient memory' run a netstat to see if its actually all the ports being blocked by a rogue program.