Testing XPSP2 system.adm -> all clients hanging

  • Thread starter Thread starter redziller
  • Start date Start date
R

redziller

Hello,

My environment is AD on 2k servers with a mix or NT4/W2K/XP & W2K
Terminal Server with Citrix XP on top.

I've bought some XPSP2 crates and am building a policy to manage the
user environment. I've created an OU with a few test users and created
a new policy with just the XPSP2 system.adm.

I use the work arround suggested in
http://support.microsoft.com/default.aspx?kbid=842933
i.e. click OK on all the error messages.

I'm testing the policy one setting at a time mainly WRT start menu /
active desktop and folder redirection. Logging on a test crate
sometimes takes a minute or so - profile is <½Mb no logon script yet,
just the policy.

But twice now I save a policy setting in the MMC where I've now snapped
in the policy (rather then edditing it off the OU) and all clients (~50
+ 15 TS sessions) hang. No server is at all busy (PDC W2K, BDC W2K,
Exchange2K on W2k, SQL2k on W2K, TS 2k & proxy on NT4).

I log into the PDC and try to open the MMC with snap-ins of AD & the
policy and it hangs for a few minutes then says it can't find a domain.

I'm then forced to "quickly" reboot (why do people ask you to do
something "quickly" as if that will then mean less work when obviously
doing the same work faster is harder?)

Nothing interesting in the event logs as far as I can see.

There is the event below common to both events but the CSV files that
the performance logger is generating are not large.

Anyone have similar experiences?

It feels like the AD is not coping with the ADM leaving clients in
limbo but there's no value in that kind of sloppy talk...





Event Type: Warning
Event Source: SysmonLog
Event Category: None
Event ID: 2025
Date: 04/03/2005
Time: 15:34:15
User: N/A
Computer: %PDC server name%
Description:
An error occurred while trying to update the log file with the current
data for the %PDC server name% log session. This log session will be
stopped. The Pdh error returned is : The event log file is full.
..
Data:
0000: de 05 00 00 Þ...
 
redziller said:
Hello,

My environment is AD on 2k servers with a mix or NT4/W2K/XP & W2K
Terminal Server with Citrix XP on top.

I've bought some XPSP2 crates and am building a policy to manage the
user environment. I've created an OU with a few test users and created
a new policy with just the XPSP2 system.adm.

I use the work arround suggested in
http://support.microsoft.com/default.aspx?kbid=842933
i.e. click OK on all the error messages.

I'm testing the policy one setting at a time mainly WRT start menu /
active desktop and folder redirection. Logging on a test crate
sometimes takes a minute or so - profile is <½Mb no logon script yet,
just the policy.

But twice now I save a policy setting in the MMC where I've now snapped
in the policy (rather then edditing it off the OU) and all clients (~50
+ 15 TS sessions) hang. No server is at all busy (PDC W2K, BDC W2K,
Exchange2K on W2k, SQL2k on W2K, TS 2k & proxy on NT4).

I log into the PDC and try to open the MMC with snap-ins of AD & the
policy and it hangs for a few minutes then says it can't find a domain.

I'm then forced to "quickly" reboot (why do people ask you to do
something "quickly" as if that will then mean less work when obviously
doing the same work faster is harder?)

Nothing interesting in the event logs as far as I can see.

There is the event below common to both events but the CSV files that
the performance logger is generating are not large.

Anyone have similar experiences?

It feels like the AD is not coping with the ADM leaving clients in
limbo but there's no value in that kind of sloppy talk...





Event Type: Warning
Event Source: SysmonLog
Event Category: None
Event ID: 2025
Date: 04/03/2005
Time: 15:34:15
User: N/A
Computer: %PDC server name%
Description:
An error occurred while trying to update the log file with the current
data for the %PDC server name% log session. This log session will be
stopped. The Pdh error returned is : The event log file is full.
.
Data:
0000: de 05 00 00 Þ...

FYI The patch now available under the KB above (842933) seems to have
resolved the issue.
 
redziller said:
FYI The patch now available under the KB above (842933) seems to have
resolved the issue.


"Seems to..." oh yeah. Not at all. The issue has since become clearer -
bad DNS settings on the PDC. I installed a Gb NIC and when setting DNS
IP I found that the earlier loop-back ip address (127.0.0.1) was
dissalowed by a local policy. So I had the PDC look at the BDC. The BDC
looks at the PDC. Bang.

So now the PDC looks at itself explicitly (ie I entered the PDC's IP
instead of loopback) whereupon it ought to look to the external DNS IPs
specified in DNS forwarding.

If only I knew Jack.

A clue was opening the policy snap-in took minutes when the DNS
settings was wrong but is instantaneous now.
 
Back
Top