Unable to reconnect to MDS process

justine.joubran · ‎03-27-2012

Hello,

I have a UCCE system composed of duplexed Proggers and an AW.

I have noticed the error Unable to reconnect to MDS process (Event ID: 32782) on the ProggerA's Event Viewer. This error is not showing on ProggerB's Event Viewer.

I have checked both Central Controllers processes but was unable to locate any errors. I have dumped the logs of MDS processes, no errors.

The mdsproc.exe_2011 ... .mdmp mini dump files all date back to 2011 on both Proggers.

The message on the Event Viewer is not significant enough to me. Can someone provide explanation or suggestion for where to gather more logs for this error?

Thanks,

Justine.

gnanem · ‎03-27-2012

Hi,

The MDS process on each ICM CallRouter sends one uni-directional Heartbeat packet every 100 ms. The MDS logs, with the proper trace bit set, notices a failed Heartbeat when the receiving side of the ICM CallRouter does not receive this packet within the designated timeout, normally set to 100 ms. The application takes corrective action only when it misses five consecutive Heartbeats.So better increse the EMSTraceMask level under Router -->EMS--Process-->MDS in regiestry and then you can able to find the this issue.

Regards,

Gnaneswaran.

justine.joubran · ‎03-27-2012

Hi,

Thanks for the explanation.

What is the value for the increased tracing level? I have it set to default value "0".

Thanks,

Justine.

gnanem · ‎03-27-2012

Hi,

You can set the trace level as 0x02000000 and after collecting logs,disable the trace level to default.

Regards,

Gnaneswaran

spaulin · ‎04-01-2012

Enabling heartbeat countdown tracing won't help you here. With all due respect to the previous poster, this is NOT a heartbeat timeout issue. This is totally and completely unrelated to any heartbeat mechanism. You do not need to enable any tracing on anything. If you already enabled MDS tracing, please disable it (EMSTraceMask = 0).

MDS doesn't connect/disconnect/reconnect to itself. Some other process couldn't connect to it.

Given that this is a "progger" this could be an issue with a PG process (likely OPC) attempting to connect to the PG's MDS, or it could be a Router or Logger (i.e. rpl) process unable to reconnect to the CallRouter service's MDS process.

You need to find the source process (what generated that event), and look at ITS process logs. Did this happen once? Or is it still happening? If it only occurred once, and it was shortly after system initialization, I wouldn't worry about it at all. If it's a constant/recurring message seen in EventVwr, then & only then would it be worth investigating.

-Steve (16+ years supporting and training on ICM/UCCE, and truly, I could go another 16 years without ever looking at Application Eventvwr logs.).

geoff · ‎04-01-2012

spaulin wrote:

-Steve (16+ years supporting and training on ICM/UCCE, and truly, I could go another 16 years without ever looking at Application Eventvwr logs.).

And one of the reasons you did not in the past is that there were a lot of useless events added to the event viewer. I certainly remember Error events being added to the event viewer that were simply notifications. It has been cleaned up a lot over the last few releases - maybe it still needs some work.

It should be an effective tool.

(PS: I love your posts)

Regards,

Geoff

justine.joubran · ‎04-01-2012

Thanks a lot for the explanations. Didn't have the chance to reply earlier.

It turned out that the related CTIOS process was in mode ACTIVATING. I had cycled the CTIOS process as the agents were connected to the other CTIOS, and the error message stoped showing.

Thanks again posters