Cisco Support for Restoring HDS Lost Data

[saharhanna]

                   Hi all,

An issue occurred that shrinked the HDS database size and a great amount of data was lost on 25/01/2013.

However, we still have the database daily backup that we can use to recuperate the data.

I can't now drop the database and restore it because I will lose the recent data that were stored in the database after the 25th.

I need to make sure if we can restore the lost data by appending to the tables the data that was lost. Does anyone know if this is supported by Cisco?

Otherwise is there any other way to perform this?

Appreciate all your answers

Thanks

Sahar

[Gergely Szabo · Accepted solution]

Hi,

not sure about Cisco's support. But theoretically, if you turn off the Loggers, and then restore the HDS database, then you should have data from the restored DB and when the Loggers wake up, they should update the HDS (provided "fresh" data is still there). Did not try it myself but if you have a 2-3 hours maintenance window, then it's worth a try.

G.

[Martin Connolly · Accepted solution]

Hi,

Don't turn off the Loggers! As the CallRouters route calls they write details to their respective Logger databases. If you turn off the Loggers for 2 hours, you'll have a black hole in your historical database in your loggers and hence your HDS of 2 hours.

Stop the ICM Distributor service on your AW/HDS

Restore your database from your backup

Once you re-start the ICM Distributor service, it will compare the highest recoverykey it has for each hds table and copy down any more recent records.

By default the Logger databases retain 14 days of historical data, so as long as your backup was not more than 14 days ago, all the new historical data from then should be copied down to the HDS.

All of this, though is predicated on you checking the following registry setting, this determines when the replication process on the Distributor starts whether it starts copying down just any new data from then or whether to copy down all the data from the last time the HDS was updated.

HKLM\SOFTWARE\Cisco Systems, Inc.\ICM\\Distributor\NodeManager\CurrentVersion\Processes\rpl\ImageArgs

This will be in the format of something like

/db /client /name /replicationport 40026 /recoveryport 40028 /all

The most important bit here is the /all at the end of the argument, this tells the replication process to copy down all historical data that it's missing.

Hope this helps.

Martin.

[saharhanna]

Hi

Thank you for the help.

We did the restore yesterday as you suggested.

We did not turn off the loggers and did not need to perform it outside working hours.

Simply we turned off Distributor and Jaguar services because they were connecting to HDS.

Then we restored the database from the backup file.

When we restarted the Distributor and Jaguar services, data was being replicated to refill all the missing dates

Thanks

Sahar