VCS Cluster - H.323 Registration

[diheldne]

Hi all,

we have an VS Cluster and would like to register 2 MCU´s and Gateway with H.323 on it.

So - registration on the first peer fine - no prob. Registration on the 2. peer - doesnt work (received fron unauthenticated source)

VCS Trace ... no crypto tokens present // no valid crypto tokens.

Its happend allways on the 2. peer.

Reboot MCU, Reboot Cluster, Check Time Servers, Disabling / Enabling H.323 in VCS and MCU - all done, still the same failure.

VCS Cluster / MCU / ISDN are all in the same subnet, but in diffrent location. RTD between the location is less than 10ms.

btw: SIP with tls is running without any isues.

Does someone have any idea ?

Many Thanks for your helping hands

/dirk

[Martin Koch · Accepted solution]

In general this should not happen.

You said you checked the time servers, but did you also check the actual time on the

VCS and the endpoint? (for example login as root and check with date).

Try to use the same NTP server on all VCS and the endpoints.

Is the clustering itself ok (see configuration / clustering)?

Anything not ok with the zone configurations?

Do you have any special local or server based CPLs on the VCS?

Any firewall/nat/alg in between the VCS and the endpoint?

Did you do a xconfig on both VCS and compare the config for differences?

Your sure its the right vcs you hit?

Did you further look into the log files and compare whats going on on the first one in comparison to the second one?

A tcpdump and debug logging on the VCSs can be interesting as well.

Please remember to rate helpful responses and identify helpful or correct answers.

[diheldne]

HI,

many thanks for your answer.  Jepp - all done, and nothing of them was really sucessfully.

The only what i found was that one certificate was on md5, the other in sha1. Normally there should be no impact, but changing both certificates to sha1 , rebooting the cluster for serveral times, rebooting the mcu - then ist works.

Dont ask me why....sometimes i it strange...