Anything in your event viewer? Usually you should get some pool
memory exhausted events. James Chong 11130 Sunrise Valley Dr. Reston, VA 20191 From: exchangelist-bounce@xxxxxxxxxxxxx
[mailto:exchangelist-bounce@xxxxxxxxxxxxx] On Behalf Of Mahadevan
Subramanian Yes Michael, Thanks for your quick response. Even i suspected the same and ran poolmon during the time when
the problem persists. But Paged and Non-Paged values are normal during that
time.... ;-( Regards... Mahadevan On 1/3/08, Michael B. Smith
<michael@xxxxxxxxxxxxxxxxxxxxxxxx>
wrote: Sounds more and more like a
non-paged-pool memory leak. Regards, Michael B. Smith MCSE/Exchange MVP http://TheEssentialExchange.com
From: exchangelist-bounce@xxxxxxxxxxxxx
[mailto:exchangelist-bounce@xxxxxxxxxxxxx]
On Behalf Of Mahadevan Subramanian
Hello James, Yeah i checked the 2080 events... As the normal topology discovery happens
every 15 min, this 2080 event should also be listed every 15 min. But once the
problem is started this event does not occur since the DS bind itself fails
with all DCs. Instead the several list of events which i have mentioned in my
previous mail gets populated. But the last topology discovery before the
problem starts (ie: before 15 min) will be succesfull and we will get a
2080 event with normal values for CDG. Regards... Mahadevan On 1/3/08, James Chong <jchong@xxxxxxxxxxxxxx> wrote: I have seen problems with
discovery failures due to SACL issues; but this would be an all or nothing deal
ie. It would work or wouldn't. Your issue is strange since it works on an off.
Is there an event id 2080. If so; can you see what values you get for
reachability and SACL and post? Event Type: Information Event
Source: MSExchangeDSAccess Event Category: Topology Event ID: 2080 Event ID 2080 from
MSExchangeDSAccess http://support.microsoft.com/kb/316300 James Chong 11130 Sunrise Valley Dr. Reston, VA 20191 From: exchangelist-bounce@xxxxxxxxxxxxx
[mailto:exchangelist-bounce@xxxxxxxxxxxxx]
On Behalf Of Mahadevan Subramanian Hello Michael, The access to the port is not actually gone. We are able to telnet the DS
servers in 389 port from the affected Exchange Server during the problem. But
rest of the tools like DCdiag, Netdiag.... etc is not running
during the time when the problem persists. Only thing is the Topology
Discovery fails to bind with any of the DS serevrs. We have got multiple DCs
which are detected in-site, once this problem happens it is trying to get the
Topology information using some boot strap server and DS bind fails with all
DCs and hence going to Out-Site DC and even that is getting failed. This does not look like a AD issue as other Exchange Servers in the
site are also using the same DCs in the DSAccess Auto Discovery mode. Any suggestions for throwing light in this dark issue is appreciated... :-) Regards... Mahadevan On 1/2/08, Michael B. Smith <michael@xxxxxxxxxxxxxxxxxxxxxxxx>
wrote: Can you "telnet
<ds-server> 389" ?? That is, is access to the port actually gone? If so, this isn't an Exchange
issue per se – the Exchange problem is just a symptom of a larger problem. Regards, Michael B. Smith MCSE/Exchange MVP http://TheEssentialExchange.com
From: exchangelist-bounce@xxxxxxxxxxxxx
[mailto:exchangelist-bounce@xxxxxxxxxxxxx]
On Behalf Of Mahadevan Subramanian Hello All, Windows 2003 AD With Native Mode Exchange 2003 SP2 with Native Mode We have vague issue in DSAccess Topology discovry only in one particular
Front End Exchange server which we use it as Mobility Server.Every week the
topology discovery will fail with 2110, 2115, 2114, 2095, 2102,
2103...... events. The main error is " Process
MAD.EXE (PID=1908). Could not bind to DS server XXXXXXXX Error 5a in Port 389".
There is no use in running any tests like netdiag, nltest, policy test during
this time because it is not able to communicate with port 389. The only
temporary slution which we have is Reboot the Exchange server after which it
will work fine for another week. Did the following things.. Increased Diagnostic Logging for DSAccess and hence got the above mentioned
events... Ran ExBPA and found nothing fishy, except 3GB switch enabled in the Front
End Server which is not there in any other FE servers. Memory usage is normal. We are planning to take this up with Microsoft but before that just thought
of getting any of your suggestions if any one has faced this issue.
|