Side A logger and side B router fails

Unanswered Question
Oct 4th, 2008

In duplex ICM environment, if logger on side A fails and router on side B fails, side B will obviously be removed from service and call routing will coninue uninterrupted on side A. But databases on both sides will be out of date as both loggers are not in service (as per SRND). When the nodes will come up, what about the data during downtime? will it be recoverable and accessible in databases?

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Average Rating: 4 (1 ratings)
Md. Mijanur Rahman Sun, 10/05/2008 - 05:12

The same tables exists on both AW (at awdb). Loggers will update soon itself from the AW database when it will be online again. But i'm not sure about the outbound tables (bA).

geoff@hp.com Sun, 10/05/2008 - 06:38

"will it be recoverable and accessible in databases?"

How could it be? Both loggers are down.

The Router has pulled the configuration into memory when it started up, so under this unusual scenario, your call center will continue to run, but historical data cannot be held in memory for long. Updates on the half-hour come from the PG, the Router saves to memory and requests the Logger to write to the database. It cannot. I'm guessing that when the next half-hour update comes, it will overwrite the previous. Data will be lost.

Regards,

Geoff

Md. Mijanur Rahman Sun, 10/05/2008 - 20:55

Yes, thanks Geoff to clear these thing, AW db shouldn't work while both loggers can't fetch data, am I right?

cvenour Sun, 10/05/2008 - 22:47

>AW db shouldn't work while both loggers can't fetch data, am I right?

The AW (rtclient process) receives it's real time data feed from the rtsvr process on the Router.

The AW (updateAW process) receives it's config data feed from the Router, after config changes are committed into (at least one of) the Loggers. No Logger = no config changes.

The HDS (replication process) receives it's feed from the replication process on the Logger. No Logger = no historical updates.

C.

Md. Mijanur Rahman Mon, 10/06/2008 - 01:38

Ok, it's now clear that Historical Database updates itself through 'replication' process exists both at AW and Logger. The process 'updateAW' is responsible to update the configuration in central database located at Logger.Whereas the 'rtclient' of AW communicates with 'dbagent' process of Router whereas the 'dbworker' process of Router access the central controller.

My question is, is 'dbagent' dependent or relate anyhow on 'dbworker' process? if the Logger goes down, can't 'dbagent' update the AW with real time feed directly? Which process is responsible to update the awdb?

cvenour Mon, 10/06/2008 - 15:31

>The process 'updateAW' is responsible to update the configuration in central database located at Logger.

No, the UpdateAW process on the AW is responsible for updating the AW's DB.

>Whereas the 'rtclient' of AW communicates with 'dbagent' process of Router whereas the 'dbworker' process of Router access the central controller.

No, the rtclient process on the AW receives Real Time data from the rtsvr process on the Router.

The dbworker process on the Router is used if you have enabled the SQL Gateway on the Router (used to make DB dips during routing scripts).

C.

Md. Mijanur Rahman Mon, 10/06/2008 - 21:20

Thanks to clear up those things :)

I had a confusion while got the message at updateaw:

'Update of central database succeeded, Message:, LastUpdateKey: '

But I also got another message:

'Successfully applied 1 transaction(s) to local database'-----it matches with your comment.

Can I have any document where I can get the relations of the processes in the ICM?

cvenour Mon, 10/06/2008 - 23:10

Cisco ICM Product Training - Part 2 - Student Guide 1 contains a complete rundown of the "significant" ICM processes.

C.

Actions

Login or Register to take actions

This Discussion

Posted October 4, 2008 at 8:31 PM
Stats:
Replies:8 Avg. Rating:4
Views:645 Votes:0
Shares:0
Tags: No tags.

Discussions Leaderboard