cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1977
Views
0
Helpful
12
Replies

Restoring a Call Manager 6.12 server on new box. How long should it take?

kwaldecker
Level 1
Level 1

Hello all,

I am doing a restore of call manager 6.1.2 onto a new server.  The restore however seems to be taking too long.

I ran it yesterday and it ran through the night.  I came back in this morning and it was still running.  So over 20 hours.  no errors on the system.  it looked like it was still running.

So i rebooted  the server and started again.  It is now 4 hours in and still going on the restore database part of it.

Any experience out there on how long it should take?

this database has 3000 phones in it and am just restoring the publisher

thanks

Ken

12 Replies 12

phooghen
Cisco Employee
Cisco Employee

The restore of the database of a publisher normally takes 4 to 6 hours.

Are you sure you followed the procedure at:

http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/install/6_1_1/cluster/clstr611.html#wp87717

Did you try :

Go to Serviceability page> Tools> Control Center> Network Services.

Restart DRF Local and DRF Master Services for each server.

.

Pierre

Hi Pierre,

will restarting those services stop the current restore process?

I'll go over that link again and see if i missed anything.

thanks for the link!

also do you know if it pulls the .tar file first then does the restore? or does it need contact with the .tar SFTP server the whole time?

the reason i ask is i have my laptop as the SFTP server and have left it at the customer site the past 2 nights and wanted to know if i can use it now that the restore has been going for awhile

thanks

I don't think the restore upload the .tar at the beginning of the restore.

In fact for each component, there is a specific .tar file (Annunciator, Bat, CCMDB, TFTP ...)

Another recommendation is to try to use another SFTP software since for example FreeFTP is limited to a file size of 1 GB.

Pierre.

thats what i was thinking too but was hoping to confirm.  I'll just leave the laptop for now.

I am using CoreFTP's  mini sftp server.  I don't see any limits on it though

I reloaded the server and ran the restore again.

this time it finished but had the following error on the screen

ERROR: Error code not found in the error map file, Restore Completed

all the data seems to be there for phones etc.

i found the error in the actual logs

here it is below

CDRREP: Could not start Cisco CAR Scheduler
CDRREP: Exiting car service activation script with error code 99
CAR services failed to start; restore failed. TERMINATING

with it being CAR do you think the rest of the restore went fine and we can move on? i don't see anymore errors

You are hitting a bug CSCsr91903 

Here is the bug detailed information:

DRS restore of CCMDB fails on new built PUB when contacting SUBs


Symptom:
During the Disaster Recovery System (DRS) restore on a newly built publisher,
the CCMDB component of the restore will results in a warning.
For 7.0(1), CUCM cluster has more than one node CCMDB logs say "No Subscriber
to be restored since it is a Single Node installation". Restore status does not
indicate any warning.

Conditions:
If the Communications Manager database needs to be restored on a freshly
installed Publisher and the Subscriber servers are currently up and
operational, the restore will not complete properly causing the re-installation
of all subscriber servers in the cluster.

Workaround:
For 6.1(2) CUCM Release
During the restore process the Publisher is not made properly aware of the
subscribers in the cluster. After the fresh installation of the Cisco Unified
Communications software, but prior to running the DRS restore, the subscriber
servers need to be added to system->server within Cisco Unified CM
Administration.
If you run into this issue install the Cisco Unified Communications Manager
software once again on the publisher server and complete the workaround above.

For 7.0(1) CUCM Release

1. Fresh install publisher
2. DRS Restore on publisher
3. Reboot publisher (CLI:utils system restart)
4. Wait for few minutes after publisher is up.
5. Restore subscribers from publisher (CLI:utils dbreplication
forcedatasyncsub all), run on publisher.
6. Reboot subscribers (CLI:utils system restart)
7. Drop admin DB on subscribers (CLI:utils dbreplication dropadmindb)
8. Drop admin DB on publisher (CLI:utils dbreplication dropadmindb)
9. Reset replication (CLI:utils dbreplication reset all)
10. Reboot publisher (CLI:utils system restart)


Further Description:
Proper way to restore a publisher without reinstalling the subscribers:
1) Fresh install Publisher with same version of the backup and current
subscribers. Reboot server.
2) All subscribers must be online. Do not shut down subscribers.
3) On the publisher
Access to CCMAdmin page through url https://:8443 then navigate to
system->server and add all existing subscribers. If DNS is not being used add
servers by IP address.
3) Navigate to DRS page from Navigation drop down proceed with restore
procedure. Don't need to activate any services prior to restore.
4) Login to the DRS pages using the platform admin user name and password
5) From DRS main page, select Backup -> Backup Device.
-Enter Device name (i.e Network_Backup) and sftp info then Click Save
6) Navigate to the Restore -> Restore Wizard.
-Select Device Name (i.e NETWORK_BACKUP) then Click Next
-Select CCM and CDR_CAR then Click Next
-Select CCM and CDR_CAR on Publisher node only then Click Restore
7) After restore has successfully completed reboot all servers in the cluster
(publisher first) navigating to Cisco Unified OS Administration and selecting
Settings -> Version -> Restart

thanks for the response!

does that bug match my error?  it seems to match the version however the only error i see in the log is related to CAR.  nothing about the subscribers

Thisbug matches the error message:

ERROR: Error code not found in the error map file, Restore Completed

gotcha,

i found the error in the log file from the real time monitoring tool as

CDRREP: Could not start Cisco CAR Scheduler
CDRREP: Exiting car service activation script with error code 99
CAR services failed to start; restore failed. TERMINATING

Hello,

I know its a little old but was wondering if you were able to get this resolved?

Thanks!

Please rate useful posts.