bjblackmore
asked on
Can not connect to Live Communications Server 2005
We're been running Microsoft Live Communications Server 2005 SP1 for the past 4 years without issue. Suddenly last week communicator clients started failing to login on all users desktops. At first it appeared that the LCS server was stopping randomly, after restarting it clients would be able to login.
However, now it appears that the server is not stopping, the service is continuing to run, there are no error message in the event logs. But when clients try to log communicator in they get an error saying:
'Cannot connect to the communications service. Please wait for your system administrator to correct this problem and try again later.'
This can happen 2 or 3 times if you retry signing in, eventually after the 3rd or 4th attempt communicator will sign in, and work correctly for an hour or so, but will then randomly sign out. Signing back in goes through the steps about with its failing then finally suceeding.
I have just performed a complete uninstall of LCS, including hotfixes, and deleted the databases, and have now reinstalled everything from the ground up, LCS 2005, then SP1, then hotfixes, and recreated the user database. However, the problem still seems to persist.
However, now it appears that the server is not stopping, the service is continuing to run, there are no error message in the event logs. But when clients try to log communicator in they get an error saying:
'Cannot connect to the communications service. Please wait for your system administrator to correct this problem and try again later.'
This can happen 2 or 3 times if you retry signing in, eventually after the 3rd or 4th attempt communicator will sign in, and work correctly for an hour or so, but will then randomly sign out. Signing back in goes through the steps about with its failing then finally suceeding.
I have just performed a complete uninstall of LCS, including hotfixes, and deleted the databases, and have now reinstalled everything from the ground up, LCS 2005, then SP1, then hotfixes, and recreated the user database. However, the problem still seems to persist.
ASKER
Hi,
Thanks for the reply. I've enabled logging, and have 3 files; Communicator0.log; LCAPI0.log & LCAPI0.log.
Which one should I be looking in, there is an awful lot of data, which doesn't kae an awful lot of sense. Do you want me to post a snip of any of the files?
Thanks for the reply. I've enabled logging, and have 3 files; Communicator0.log; LCAPI0.log & LCAPI0.log.
Which one should I be looking in, there is an awful lot of data, which doesn't kae an awful lot of sense. Do you want me to post a snip of any of the files?
ASKER
Looking in LCAPI0.log the only error I can find is:
09:12:43.002 D28:1654 ERROR :: ASYNC_SOCKET::OnTimerExpir
09:12:43.002 D28:1654 ERROR :: OUTGOING_TRANSACTION::OnRe
09:12:43.002 D28:1654 INFO :: CRTCProfile::SetState - [0185A2F8] enState:1->5, statusCode=80ee001c, text=(null)
09:12:43.002 D28:1654 INFO :: RegistrationStateChangeEve
09:12:43.002 D28:1654 INFO :: CRTCProfile::SetState - register error, clear postponed watchers
Not sure what this means, why its eventually successful after 3 or 4 attempts, or how to fix it!
Ben
09:12:43.002 D28:1654 ERROR :: ASYNC_SOCKET::OnTimerExpir
09:12:43.002 D28:1654 ERROR :: OUTGOING_TRANSACTION::OnRe
09:12:43.002 D28:1654 INFO :: CRTCProfile::SetState - [0185A2F8] enState:1->5, statusCode=80ee001c, text=(null)
09:12:43.002 D28:1654 INFO :: RegistrationStateChangeEve
09:12:43.002 D28:1654 INFO :: CRTCProfile::SetState - register error, clear postponed watchers
Not sure what this means, why its eventually successful after 3 or 4 attempts, or how to fix it!
Ben
fair point - there is alot data, could you perhaps zip all of them up and attach ?
the error will be in there somewhere - like you say, its just a matter of finding it!
the error will be in there somewhere - like you say, its just a matter of finding it!
ASKER
Hi,
Here are the 3 files zipped up.
To try & keep the file size small, this is a trace from opening Communicator and trying to login, and it failing a single time. I can supply a file from multiple failures, or from a successful login if required.
tracing.zip
Here are the 3 files zipped up.
To try & keep the file size small, this is a trace from opening Communicator and trying to login, and it failing a single time. I can supply a file from multiple failures, or from a successful login if required.
tracing.zip
hurumph - i see what you mean
so - _sip._tls.ourdomain.com is defined as 5061 pointing to exchange.ourdomain.com - and you can telnet to that ok on 5061 from a client expieriencing the issue ?
there seems to be communications troubles with 192.168.254.2:5061 - which i assume is "exchange.ourdomain.com " - flat out, if you can telnet to that address - there must be service down or something busted on the server - in which case, the OCS event log is the place to start.
so - _sip._tls.ourdomain.com is defined as 5061 pointing to exchange.ourdomain.com - and you can telnet to that ok on 5061 from a client expieriencing the issue ?
there seems to be communications troubles with 192.168.254.2:5061 - which i assume is "exchange.ourdomain.com " - flat out, if you can telnet to that address - there must be service down or something busted on the server - in which case, the OCS event log is the place to start.
ASKER
Hi,
Thanks again for the reply.
I can't tell net to exchange.ourdomain.com:506
I've set the LCS logging level to 4, and tried to sign in, attached is the log, I tried to keep it small, as it seems to be very details.
I thought I'd fixed the problem a while ago, after posting the following question https://www.experts-exchange.com/questions/24630718/DNS-Using-Nearly-Every-Well-Known-Port.html?anchorAnswerId=25031768#a25031768 which turned out to be DNS grabbing a low ephemeral port range. However, after fixing that DNS now grabs ports 49,100 and above, which shouldn't be interfering with LCS. And even disabling the DNS Server service doesn't fix ths issue, so I don't think its DNS that s causing the problem.
exchange.ourdomain.com is the server LCS is running on, it used to host exchange 2003, hense the name, but that has now been placed on a newer server, however, the name was never changed on this server. The network card has 3 IP addresses assigned, 192.168.254.2, .3 & .4
lcssiplog-2009080612-001.log
Thanks again for the reply.
I can't tell net to exchange.ourdomain.com:506
I've set the LCS logging level to 4, and tried to sign in, attached is the log, I tried to keep it small, as it seems to be very details.
I thought I'd fixed the problem a while ago, after posting the following question https://www.experts-exchange.com/questions/24630718/DNS-Using-Nearly-Every-Well-Known-Port.html?anchorAnswerId=25031768#a25031768 which turned out to be DNS grabbing a low ephemeral port range. However, after fixing that DNS now grabs ports 49,100 and above, which shouldn't be interfering with LCS. And even disabling the DNS Server service doesn't fix ths issue, so I don't think its DNS that s causing the problem.
exchange.ourdomain.com is the server LCS is running on, it used to host exchange 2003, hense the name, but that has now been placed on a newer server, however, the name was never changed on this server. The network card has 3 IP addresses assigned, 192.168.254.2, .3 & .4
lcssiplog-2009080612-001.log
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
cool - but may i ask why, if your starting from scratch, your going with LCS instead of OCS ?
ASKER
Few reasons, the existing server is 6 years old, speed to reinstall LCS - I managed to reinstall in under 2 hours. Plus we're only a small business with 16 employees, don't need the extra functionality OCS would have offered right now.
hopefully it will give us some more detail as to what part is actually failing.