Hello,
SeveralDP clients shouldmove fromsite Ato siteB. So they will change Cell Manager (CMa to CMb).
On site A : CMa is DP 7.01 on Windows 2008 R2 - CMa is MA
On site B : CMb is DP 7.01 on Windows 2008 R2 - MA is another server.
What shouldI do ?
1.uninstall DPclients from theCMa thenre-install from theCMb ?
2.another possiblesolution ?
I willalso change thesetting OB2PORTRANGESPEC in theomnircfile for eachclient becausetheport rangeis notthe same.
Thanks for your help.
Stephane
Hi,
run this command and post output:
omnidbuitl -info
omnidbutil -extendinfo
These commandswill give usan idea of thestate of yourdatabase
Regards
Hi,
One simple step is to rename the regedit, and this does not have to uninstall anything.
example:
In one client CMa, you go to regedit and delete the name in the site(delete CMa name)
Go to: Start / Run: regedit
HKEY_LOCAL_MACHINE \ SOFTWARE \ Hewlett-Packard \ OpenView \ OmniBackII \ Site \ CellServer (Data = Cell Manager name)
In the name you can put the CMb and then go to the GUI on CMb and export this client.
If it does not work ,just delete the name of CMa and export it in the GUI CMb that client and it should work.
And with OB2PORTRANGESPEC you only change the corresponding values
I hope this helps.
My experience has been that having more than 20 hosts in a backup specification causes problems.
It may be different for you, I'm on quite old hardware.
Jeremy
Hi Eblyn,
Thanks for your answer.
So before relocation i have to :
1. on client : change CM name in regedit for Windows clients (but for RedHat, AIX and Debian ?)
2. on CMa : export client (in GUI right-click client then click Delete without uninstall DP)
After relocation i have to :
1. on CMb : import client
2. on client : modify OB2PORTRANGESPEC in omnirc file
Is this ok ?
Hi,
GRE is installed / started as a separate program on the Exchange server. Once you install the component on the server you should have a start menu item refering to it....
Regards,
Nikhil
1. On Unix/Linux the file is /etc/opt/omni/client/cell_server
There can be several reasons....but 24 hours sounds a bit too much
Things you could check if it is Windows:
- Virus Scanner
- Any monitoring processes (I recently encountered a similar problem caused by HP System Monitor hpshmd.exe slowing down the backups to almost crawling speed)
- Firewall
- DNS resolution etc. etc.
If it is happening only on weekends or particular days of the week, then it might be worthwhile to check the Windows scheduler too.
On Unix, you could run glance (or similar performance monitor on Linux) to check if there are any network blockages, ping issues etc. Check crontab if the problem is seen on particular weekdays.
As strange it might sound, but the problem can be on the CM itself. If it is a large filesystem with lot of objects that you are backing up, there are many IDB entries to be done...and if the IDB maintainance kicks in at the same time, the CM system might be stressed...
If the startup is being delayed, I rather suspect something related to CM, because if the problem was on client site, the backup would have aborted / failed due to timeout....
So in short, there are too many variables for your problem, try to isolate by eliminating one by one.
Regards,
Nikhil
I am attempting to restore a VMDK to a local drive to mount and get a file from it. I get the following error about 40-50GB into the restore:
[Major]
From: RSM@<name>.com "" Time: 2/11/2014 1:25:44 PM
[
61:3003] Lost connection to OB2BAR restore DA named ""
on host <name>.com.
Ipc subsystem reports: "IPC Read Error
System error: [10054] Connection reset by peer
I am running DP 7.03 on Server 2008R2. I have restored VMs in the past, but no VMDKs this big. I suspect I may need to adjust something in the omnirc file?
Any insights would be appreciated.
Thanks!
Brian
Hi,
about your questions
So before relocation i have to :
1. on client : change CM name in regedit for Windows clients (but for RedHat, AIX and Debian ?)
nickil answer: . On Unix/Linux the file is /etc/opt/omni/client/cell_server , delete or rename this file, afther that import the client on new CM
2. on CMa : export client (in GUI right-click client then click Delete without uninstall DP), R/Yes, is correct.
After relocation i have to :
1. on CMb : import client R/Yes.
2. on client : modify OB2PORTRANGESPEC in omnirc file R/Yes.
Regards
**Clicking the KUDOS star on the left is a way to say 'Thanks'**
Sorry Eblyn...you are correct....my reply was slightly inadequate...one needs to delete/move the file. Like Windows you can just change the contents to reflect the new CM name, it is just a plain text file with the CM name (FQDN).
After that you would need to check the installation. On many Linux distributions rexec/rlogin are not installed, so in case the installation server has also changed, then you need to copy the SSH keys from the old IS to new IS and enable SSH in omnirc to perform any install/check/upgrade operations.
I would personally prefer to MOVE the file (to say /tmp) just to avoid confusion and import the host to the new CM.
Cheers,
Nikhil
The first 4 suggestions check out okay.
What i can add is that this backup job that is delayed so long is a fileserver, so it has a large number of small files. However there are multiple file servers and the other ones don't seem to have this issue.
Can you check the logging level for that object? In case of large number of small files changing the logging level to "Directories Only" would help.
Work´s with Exchange 2013???
There is nowhere near enough information for me to make even the most basic evaluation.
What version of DP?
What is the OS-type and version of the Cell Manager OS?
were there any other warnings or errors prior to this?
"Connection reset by Peer" usually indicate a network disconnect, or it may be the result of a pevious error. if it is an error seemingly associated with any other type of event, you may want to review the attached document
I would pay special attention to item # 10, to increase the TcpMaxDataRetransmissions registry parameter
This article was written back in the day of Windows NT/4, but, it is just as valid today, the one difference being, you may not have to reboot the server to make this change effective
Hello,
We are on DP A 0.7.00 Build 105 and have been using com encryption + software encryption without compression on an LTO4 drive using LTO3 media. We recently added an LTO6 library to our infrastructure and would like to move to drive based encryption.
1) What things do we need to lookout for in moving to drive based encryption
2) Will we need to generate new keys or an additional key or can we use the ones in place for each backup object?
3) Is there a whitepaper / tutorial for what settings are needed to make this change?
4) After moving to drive based enc, will we have anything to look out for in restoring our old data?
5) Can you use LTO3 media for drive based encryption as long as you are using LTO4+ drives?
6) Can we do encryption and compression simultaneously with drive based encryption and actually get some gains in storage on LTO6 media?
7) Circling back around to #2 -- I've seen many posts saying drive based encryption makes a new key for each tape (at some point in the past). Has this been fixed now to be more sane?
Thanks in advance.
Hello,
I tested StoreOnce Software Catalyst Store with SAP R/3 integration. Server Side and Target Side deduplication is working fine, but Source Side deduplication can not start.
After starting backup session a receive message: There are no available connections on store 'D2D_Soft' (through gateway 'D2D_Soft_Source_side').
I built a similar scheme using StoreOnce VSA - Source Side deduplication for SAP R/3 Application is working.
Source Side deduplication with DP StoreOnce Software Catalyst Store for SAP R/3 client working or not in Data Protector (any release)?
Anyone tried it like that?
1) What things do we need to lookout for in moving to drive based encryption
Keystore file corruption. However the Keystore is now included in the IDB backup, so that, if you backup the IDB every day, you should be OK
For optimal performance, the block size used should be at least 256 kilobytes
2) Will we need to generate new keys or an additional key or can we use the ones in place for each backup object?
7) Circling back around to #2 -- I've seen many posts saying drive based encryption makes a new key for each tape (at some point in the past). Has this been fixed now to be more sane?
You should be able to use the same keys as the ones in place for each backup object
My impression is that you can create a new encryption key for each media, but this is optional. I have looked at a lot of keystore files, and, with hundreds of media still under prote4ction, I saw nowhere near the corresponding number of Encryption keys
3) Is there a whitepaper / tutorial for what settings are needed to make this change?
None that I can find, but, in the GUI, click on Help -> Topics, and search for "drive based encryption", it will give you some additional information. Also, if you click on Help -> GUides, find the CLI (Command Line Interpreted) Guide, in section 1M, get the usage for 'omnikeytool'
4) After moving to drive based enc, will we have anything to look out for in restoring our old data?
I think that this is covered in the on-line helpl, mentioned in the last response
5) Can you use LTO3 media for drive based encryption as long as you are using LTO4+ drives?
This is probably not a question that we can address from a Data Protector perspective... rather, it should be taken up with your hardware vendor
6) Can we do encryption and compression simultaneously with drive based encryption and actually get some gains in storage on LTO6 media?
First, NEVER use Software compression
Again, this is not something that we can answer from the Data Protector side of the house
