02-01-2012 04:28 AM
Hi catostropic monstrous events going on.
I need to restore some exchange databases, which are disk images on local drives on my master server. Dont need to go to tape.
The server has been rebuilt, and has no client installed on it, and I do not have an option to install it at present.
what does one do when trieng to restore data for clients that no longer exist?
Is an exchange netbackup client a different isntall than the regular client ?
How can I point to an image and just restore it to a location of my choice? the backup and restore client on the master server asks for a rsetore client and type....
from the command line...
bpimagelist -d 01/27/2012 -policy mbx1-store -U
No entity was found.
Any help would be extremely appreciated !
02-01-2012 05:04 AM
If you are trying to restore the exchange and if it is crashed you might be restoring it on top of it, or you can restore to recovery storage group in exchange.
If it was the flat database backup then you can restore it to any other server that has client installed. Do the restore from any of the media or master server.
In the BAR, select the source client as original server
Destination client where you want to restore
Policy type of restore Ms-exchange server or Ms-windows-nt if it was a flat backup as windows
Client is same its only the selection is different if you are restorig the exchange database.
02-01-2012 05:05 AM
A Netbackup client install contain the Exchange database agent from version 6.5 and up.
You can always restore data even if the original client is gone. It's just a question of entering the right names and giving the right permissions (aka altnames)
Please refer to the Netbackup Exchange admin manual how to perform alternate restores (chapter 6):
http://www.symantec.com/docs/DOC3669
PS: Image list command, is mbx-store the configured name in Netbackup ?
bpimagelist -d 01/01/2011 -client acme
and
bpimagelist -d 01/01/2011 -client acme.domain.com
will not return the same information
02-01-2012 05:21 AM
Hi thanks for the replies - I get:
ERROR: valid archive image produced, but no files deleted due to no fatal problems
I Imagine this error message is not the sharpest prose....?
Any ideas?
how can I restore from the image I have found in the cataloge?
Import it? then how do I get it to disk?
02-01-2012 05:55 AM
A archive read data and delete files on the source
A Backup read data but does not delete files.
Are you trying to run an archive ?. Have you verifyed the images are in the catalog ?
find the images file on the disk, take note of the file name. It shoud be in the format of hostname_1234567:
02-01-2012 06:11 AM
Yes I can see them in the cataloge and know exactly where the image file is and the filename.
I just dont know how to do the restore outside of "backup and restore" which is giving me the error above ;)
02-01-2012 06:15 AM
02-01-2012 06:21 AM
02-01-2012 07:32 AM
You can only restore Exchange Data to an Exchange Server with a NetBackup client installed
Set up a Recovery Storage Group on a working Exchange Server of the same Exchange Type and restore it there using the original client as the source and the new one as the destination
All done in the Backup, Archive and Restore GUI - run this from the Master Server to overcome the alternate client restore issues
See the Admin Guide for full details
02-01-2012 08:06 AM
Thanks mark - I have the client installed on a second exchange server now.
and the correct public folder backup selected for restore. Just waiting to restart the new exchagne server.
ONe more question - when these are restored to a recovery store - are the USN numbers updated so that the data appears newer that the "bad" public folders we have out there?
02-01-2012 08:25 AM
The RSG is pretty isolated
If you want this to be restored for Production use then overwrite the original (if you are sure the backup you have is good)
Your decision - may be safe just to restore to a RSG and then extract what you need from it using Exchange tools
02-02-2012 04:58 AM
Its ok - we are going to use an isolated environment on a new server. RSG doesnt work with public folders - which we lost to replication. everything is cool
02-02-2012 05:34 AM
Great - when all is solved dont forget to come back and mark the post that helped the most as a solution to assist others in the future with a similar issue
02-03-2012 12:13 AM
Argh it is far from solved it has one of two errors now.
We changed the IP address of the master to put it in a sandboxed environment, with one exhcnage server(that we are trying to restore to), and one domain controller.(required by exchange)
The master can see the client (same name as original box) but it kept failing with an error on restore status code 205: cannot connect to server backup restore manager
I thought as this was my first attempt at a restore from staging disk, maybe that was the issue - I got the same result from one that had been duplicated to tape a few days earlier.
This is for a public folder store, and we had it dismounted, and set to be overwritten as per the various documents.
I have had 3 symantec engeneers try to figure this out, to no avail - All connectivity and command line tests came up clean. Telnetted to various ports etc...
I then re-enabled the second NIC, ont the original server that that it could talk to the rest of the clients, and dropped the gate way on the sandboxed vlan. Full connectivity restored.
From here I installed Kroll Ontrack power controls - latest version and set it up to emaulate an exchange 2007 server, and followed their documentation to the letter. That server ixbackup-lon has the netbackup client installed - and can contact the master.
The Master server is not even making an attempt to contact the client - it doesnt get that far. So whether is an emulated exchange server (Kroll) or a real server, it just doesnt get that far in the restore process. Ive been using netstat on both ends, looking for established connections - there are non between client and master server (also media server)
This is Master server tape backup -> Ontrack Emulated Exchange server:
02/02/2012 23:39:09 - begin Restore 02/02/2012 23:39:12 - 1 images required 02/02/2012 23:39:12 - media CLU223 required 02/02/2012 23:39:12 - media CLU231 required 02/02/2012 23:42:06 - restoring image mbx1-lon_1327587040 02/02/2012 23:58:07 - end Restore; elapsed time: 00:18:58 cannot connect on socket(25) 03/02/2012 00:13:05 - requesting resource CLU223 03/02/2012 00:13:05 - awaiting resource CLU223 Reason: Tape media server is not active, Media Server: potter, Robot Number: 0, Robot Type: TLD, Media ID: N/A, Drive Name: N/A, Volume Pool: N/A, Storage Unit: N/A, Drive Scan Host: N/A 03/02/2012 00:18:06 - Error bpbrm(pid=3040) bpcd on ixbackup-lon exited with status 59: access to the client was not allowed 03/02/2012 00:18:07 - connecting
This is the master server Disk Backup (staging) -> ontrack emulated Exchange server:
02/02/2012 23:10:49 - begin Restore
02/02/2012 23:10:52 - restoring image mbx1-lon_1327731028
02/02/2012 23:16:09 - Error bpbrm(pid=3268) bpcd on ixbackup-lon exited with status 59: access to the client was not allowed
02/02/2012 23:16:15 - Error bpbrm(pid=3268) client restore EXIT STATUS 59: access to the client was not allowed
02/02/2012 23:27:04 - restored image mbx1-lon_1327731028 - (cannot connect on socket(25)); restore time 00:16:12
02/02/2012 23:27:05 - end Restore; elapsed time: 00:16:16
the restore failed to recover the requested files(5)
From Master server -> newly built Exchange server with same name - but on different IP address
02/02/2012 19:25:05 - begin Restore 02/02/2012 19:25:07 - restoring image mbx1-lon_1327731028 02/02/2012 19:26:36 - end Restore; elapsed time: 00:01:31 the restore failed to recover the requested files(5)
I am most concerned with restoring the the emulated Ontrack exchange environment, as we are now back on the prodcution vlan ready for the weekends backups. Any ideas?
From IXBACKUP-LON - where I want to restore from:
This is the active conenctions at start of restore when initiated from ixbackup - harry is the master server
TCP 10.1.2.186:13724 harry:2861 ESTABLISHED TCP 10.1.2.186:13724 harry:2862 ESTABLISHED TCP 127.0.0.1:59888 IXBACKUP-LON:60176 ESTABLISHED TCP 127.0.0.1:60176 IXBACKUP-LON:59888 ESTABLISHED TCP [::1]:59884 IXBACKUP-LON:59885 ESTABLISHED TCP [::1]:59885 IXBACKUP-LON:59884 ESTABLISHED
02-03-2012 12:26 AM
Back to basics.
Your client cannot resolve the master's hostname to the SERVER entry.
Add hosts entries for master and client to ensure correct forward and reverse lookup using original hostnames.
02-03-2012 12:42 AM
Hi Marianne - I have added host entries on both clients to be sure.
I can ping, telnet both ways.
I enabled verbose debug logging on the client and the errors I get are for our old media server trying to connect, rather than the master. - strange as the backup data is on the master's local disks, nothing to do with the old media server.
ixbackup-lon does is not listed in the "clients" section of the admin console - however I can query it etc
This is the error in the activity log:
03/02/2012 07:54:50 - begin Restore 03/02/2012 07:54:53 - restoring image mbx1-lon_1327731028 03/02/2012 08:00:11 - Error bpbrm(pid=5840) bpcd on ixbackup-lon exited with status 59: access to the client was not allowed 03/02/2012 08:00:17 - Error bpbrm(pid=5840) client restore EXIT STATUS 59: access to the client was not allowed 03/02/2012 08:11:06 - restored image mbx1-lon_1327731028 - (cannot connect on socket(25)); restore time 00:16:13 03/02/2012 08:11:12 - end Restore; elapsed time: 00:16:22 the restore failed to recover the requested files(5)
AND FROM THE CLIENT BPCD verbose:
08:00:10.757 [2068.1672] <2> bpcd main: VERBOSE = 5
02-03-2012 12:46 AM
Add SERVER entry on the client for: potter.internal.robertwalters.com
It seems the SERVER entry on the client is 'harry'.
: Connection from host potter.internal.robertwalters.com (10.2.20.20) port 4896
02-03-2012 12:50 AM
Harry is the master server and the media server with the data on it. Potter is an old server with a single tape drive attached, that we occasionaly use. IT did however crash a few days ago.
I need to keep the master as harry - but somehow add potter ? Potter is a little sick ;)
This is me starting potter services 5 minutes after the last restore attempt that the logs are from:
Starting services > NetBackup Client Service > NetBackup Client Service -- STARTED > NetBackup SAN Client Fibre Transport Service > NetBackup SAN Client Fibre Transport Service -- STARTED > NetBackup Event Manager > NetBackup Event Manager -- DISABLED > NetBackup Volume Manager > NetBackup Volume Manager -- STARTED > NetBackup Device Manager > NetBackup Device Manager -- STARTED > NetBackup Remote Manager and Monitor Service > NetBackup Remote Manager and Monitor Service -- STARTED > NetBackup Compatibility Service > NetBackup Compatibility Service -- STARTED > NetBackup Request Daemon > NetBackup Request Daemon -- STARTED > NetBackup Job Manager > NetBackup Job Manager -- DISABLED > NetBackup Policy Execution Manager > NetBackup Policy Execution Manager -- DISABLED > NetBackup Service Layer > NetBackup Service Layer -- STARTED > NetBackup Vault Manager > NetBackup Vault Manager -- STARTED > NetBackup Storage Lifecycle Manager > NetBackup Storage Lifecycle Manager -- NOT STARTED > NetBackup Service Monitor > NetBackup Service Monitor -- STARTED > NetBackup Bare Metal Restore Boot Server > NetBackup Bare Metal Restore Boot Server -- STARTED Start up of one or more services failed.
02-03-2012 12:53 AM
I am still unsure as to why potter is trying to get involved - as the backup and the restore and the master are all harry. Maybe I need to change something on the master properties on the server?
is it worrying that ixbackup-lon is not listed as a client? or does it only get in this list if it has at least one backup job, or belongs to a policy?
02-03-2012 12:55 AM
The client is resolving the master server connection attempt to that name.
Do the following - Clear host_cache on master and client:
bpclntcmd -clear_host_cache
Then test connectivity from master to client:
bptestbpcd -client <hostname> -verbose
Check client's bpcd log file.