03-22-2023 06:34 AM
hello,
We have a six years Veritas NetBackup solution. (about 1.5TB catalog size).
We need to enable accelerator in order to reduce our backup time.
1 - When enabled, we do a backup of a 60GB VM (windows) .
it's take 1:25 hour for the first backup.
But during the second backup it's take the same time.
2 - In job detail, we also see that netbackup is backing up ... files !!! and not working at block level.
is it a lies that netbackup is working in block mode (BLIB) ?
Solved! Go to Solution.
03-24-2023 12:33 AM
in order to sum up the troubleshooting.
What have been done :
03-22-2023 06:53 AM
In job detail, we also see that netbackup is backing up ... files !!! and not working at block level.
Sounds wrong, how did you configure the policy ? Pls post a screen dump or similar.
03-22-2023 07:12 AM
Policy Type : VMWare
Data classification : <NO data class...>
Policy storage : OST-001 (DXi Quantum , with optimized flag enable)
USe accelerator : checked
(all other option are greyed out)
Go into effect at : date ok
------------------------------------------------------------------------------------------
Schedule tab :
WeeK : Type full backup (force rescan disable)
-------------------------------------------------------------------------------------------
Clienst tab :
Select manualy, VM_name and Hardware vmx-13, OS : Windows8server64guest)
------------------------------------------------------------------------------------------
Backup selections:
ALL_LOCAL_DRIVES ... (but wth ... speaking of drive ... but for block mode ?)
-------------------------------------------------------------------------------------------
VMWare tab:
VMWare backup host : Backup MEdia server
Optimizations : enable => Enable file recovery from VM backup, Enable block-level incrmental backup (greyed out), Exclude deleted blocks, Exclude swap and paging files.
Primary VM id. : VM display name
Existing snapshot handling:
Remove Netbackup snapshot and continue backup
Transport modes : san (mode)
Advanced button :
Virtual machine quiesce : enable
Virtual disk selection : include all disk
03-22-2023 08:21 AM
Hello,
how have you determined that accelerator is not working? Even during accelerator based backups, also modified files are backed up + catalogued.
Check Accelerator Ratio in Activity Monitor column if it is really 0%. If yes, pls paste the whole Job Detail here.
Regards
M.
03-22-2023 08:28 AM - edited 03-22-2023 08:30 AM
There is no problem with this.
Netbackup is cataloging VM files during backup. If you disable "enable file recovery from VM", you will not see the files.
And do not bother with "all local drivers" at the policy. It is okay.
please post the "detailed status" of a vm backup job
03-23-2023 01:58 AM
here is the job detail :
22 march 2023 15:58:55 - Info nbjm (pid=6604) starting backup job (jobid=544876) for client xxxxxxxx, policy TEST_CBT, schedule WE
22 march 2023 15:58:56 - estimated 16954084 kbytes needed
22 march 2023 15:58:56 - Info nbjm (pid=6604) started backup (backupid=xxxxxxxxx_xxxxx) job for client xxxxxx, policy TEST_CBT, schedule WE on storage unit OST-001 using backup host xxxxxx
22 march 2023 15:58:58 - started process bpbrm (pid=10568)
22 march 2023 15:59:00 - Info bpbrm (pid=10568) xxxxxxxx is the host to backup data from
22 march 2023 15:59:00 - Info bpbrm (pid=10568) reading file list for client.
22 march 2023 15:59:01 - Info bpbrm (pid=10568) accelerator enabled
22 march 2023 15:59:02 - connecting
22 march 2023 15:59:02 - connected; connect time: 0:00:00
22 march 2023 15:59:03 - Info bpbrm (pid=10568) starting bpbkar32 on client
22 march 2023 15:59:03 - Info bpbkar32 (pid=7236) Backup started
22 march 2023 15:59:03 - Info bpbkar32 (pid=7236) accelerator enabled backup, archive bit processing:<disabled>
22 march 2023 15:59:03 - Info bptm (pid=7076) start
22 march 2023 15:59:03 - Info bptm (pid=7076) using 1048576 data buffer size
22 march 2023 15:59:03 - Info bptm (pid=7076) using 32 data buffers
22 march 2023 15:59:05 - Info bptm (pid=7076) start backup
22 march 2023 15:59:05 - Info bpbkar32 (pid=7236) INF - Backing up vCenter server xxxxxx, ESX host xxxxxxxxxxxxxxxxxxxxxxx, BIOS UUID xxxxxxxxxxxxx, Instance UUID xxxxxxxxxxxxxxxxxxxx, Display Name xxxxxxxxxx, Hostname xxxxxxxxxxxxxxxxxxx
22 march 2023 15:59:07 - begin writing
22 march 2023 16:00:18 - Info bpbkar32 (pid=7236) 0 entries sent to bpdbm
22 march 2023 16:07:15 - Info bpbkar32 (pid=7236) 95000 entries sent to bpdbm
22 march 2023 16:14:13 - Info bpbkar32 (pid=7236) 190000 entries sent to bpdbm
22 march 2023 16:23:12 - Info bpbkar32 (pid=7236) 285000 entries sent to bpdbm
22 march 2023 16:28:58 - Info bpbkar32 (pid=7236) 340762 entries sent to bpdbm
22 march 2023 16:28:59 - Info bpbkar32 (pid=7236) 340763 entries sent to bpdbm
22 march 2023 16:29:00 - Info bpbkar32 (pid=7236) 340796 entries sent to bpdbm
22 march 2023 16:29:01 - Info bpbkar32 (pid=7236) 340797 entries sent to bpdbm
22 march 2023 16:29:04 - Info bpbkar32 (pid=7236) 341068 entries sent to bpdbm
22 march 2023 16:29:04 - Info bpbkar32 (pid=7236) INF - Transport Type = nbd
22 march 2023 16:29:15 - Info bpbkar32 (pid=7236) 341128 entries sent to bpdbm
22 march 2023 16:29:16 - Info bpbkar32 (pid=7236) 341129 entries sent to bpdbm
22 march 2023 16:29:16 - Info bpbkar32 (pid=7236) accelerator sent 160603648 bytes out of 14435973120 bytes to server, optimization 98.9%
22 march 2023 16:29:18 - Info bpbkar32 (pid=7236) bpbkar waited 0 times for empty buffer, delayed 0 times.
22 march 2023 16:31:22 - Info bptm (pid=7076) waited for full buffer 154 times, delayed 115739 times
22 march 2023 16:31:28 - Info bptm (pid=7076) EXITING with status 0 <----------
22 march 2023 16:31:29 - Info bpbrm (pid=10568) validating image for client xxxxxxx
22 march 2023 16:31:31 - Info bpbkar32 (pid=7236) done. status: 0: the requested operation was successfully completed
22 march 2023 16:31:31 - end writing; write time: 0:32:24
the requested operation was successfully completed (0)
03-23-2023 02:23 AM
I do not find any problems at this job. Accelerator is enabled and sends only 1% of the data to the backup server.
This job was finished in 0:32:24 and not in one and a half hours.
What is your expectation?
03-23-2023 02:38 AM
when re do the same backup twice ... 32 mn is long.
If we compare with veeam (with the same VM 60GB) ... it's only take 4 mn.
I'm skeptical that netbackup actually uses vmware's CBT technology.
Because when you redo the same backup without adding data to the VM, the backup time must be close to a few minutes.
03-23-2023 02:45 AM
ok
Do a try. Disable all of the following or just the "enable file recovery from VM backup"
and run an full with accelerator force rescan. Then run a backup without the rescan.
Post the time of the second backup
03-23-2023 02:52 AM
thank you for your help stefanosM.
1 - i will test those options (disabled), and come back here
2 - full rescan etc, we already test that but with no effect. (before). But i must admit that with transport mode to "nbd" , we achieve to reduce from 1h30 to 32mn.
03-23-2023 03:14 AM
3 - the thing that i do not understand is the option "Exclude swap and paging files" : why this option ? if netbackup is really working at block level ? (this option mean the contrary, that the solution look into file level ... so not working at block level)
03-23-2023 03:30 AM - edited 03-23-2023 03:43 AM
Netbackup is the only backup software that can read the block stream and understand what is inside. This future is called flashbackup, and it has been part of NetBackup for 20+ years now. Initially, it was and is used for disk block backups.
This future will give you the unique opportunity to catalog the files inside the VM and exclude page files or deleted blocks. Also gives you the ability to restore a file not only to a VM but effectively everywhere. To a physical system, for example.
And you can search for a file even you do not know when it was deleted, without having to mount backup images.
The only drawback is that the backup is slower than without this option. If this is worth it or not, that is up to you.
03-23-2023 03:40 AM
To add to this. These are enhancements that can be disabled if needed but help tremendously. I'm running FlexScale 3.0.0.1 and Netbackup 10.0.
Something is not configured correctly on either the Netbackup side or the VMware side. I have a 500GB VM. With Accelerator enabled running over NBD to a Veritas FlexScale appliance. It does a full backup in ~7 minutes. Screenshot from OpsCenter attached because this backup was from last week.
I've also included a screenshot from a 3TB VM that completed in just under 10 minutes with the same configuration options as the example above. I can with 100% certainty say that accelerator does indeed work.
03-23-2023 06:13 AM
we have netbackup 7.7.3 , do you think it's a part of the problem ?
03-23-2023 06:25 AM
Is there an how to or documentation showing all the parameters needed for creating a policy for a full backup taking advantage of CBT , in order to have the same performance of veeam backup. ?
03-24-2023 12:33 AM
in order to sum up the troubleshooting.
What have been done :
03-29-2023 11:20 PM
Hello,
I don't see any point why the previous post has been marked as a solution.
Regards
M.