10-21-2011 10:22 AM
Good afternoon,
I'm currently having an issue where windows server 2k8 VM's are backing up but not restoring using the vstorage API. XP/2k3 seem fine. I have tried 3 different machines and all have given me the same result. Also a process is locking the newly restored disk files from being able to be deleted so now I'm stuck with a whole bunch of "ghosted" vmdk files that I cannot remove without killing the processes through ssh.
I'm running Netbackup 7.1.0.1 on both the master and client.
Any ideas?
Thanks in advance,
Tim
Solved! Go to Solution.
11-22-2011 08:02 AM
Upgrading to 7.1.0.2 did the trick. Thanks for everyone's help on this.
10-21-2011 10:27 AM
O yeah here are the logs on the netbackup side.
10/21/2011 12:05:42 PM - begin Restore
10/21/2011 12:05:46 PM - media %MediaID% required
10/21/2011 12:05:46 PM - restoring image %VcenterName%_1319156026
10/21/2011 12:05:49 PM - Info bptm(pid=4328) start
10/21/2011 12:05:49 PM - started process bptm (4328)
10/21/2011 12:05:49 PM - Info bpdm(pid=4328) reading backup image
10/21/2011 12:05:49 PM - Info bptm(pid=4328) using 30 data buffers
10/21/2011 12:05:49 PM - Info bptm(pid=4328) spawning a child process
10/21/2011 12:05:49 PM - Info bptm(pid=4328) child pid: 6528
10/21/2011 12:05:49 PM - Info bptm(pid=6528) start
10/21/2011 12:05:49 PM - started process bptm (6528)
10/21/2011 12:05:49 PM - Info bptm(pid=4328) Waiting for mount of media id %MediaID% (copy 1) on server %MasterServer%.
10/21/2011 12:05:49 PM - started process bptm (4328)
10/21/2011 12:05:49 PM - mounting %MediaID%
10/21/2011 12:05:49 PM - Info bptm(pid=4328) INF - Waiting for mount of media id %MediaID% on server %MasterServer% for reading.
10/21/2011 12:05:49 PM - requesting resource %MediaID%
10/21/2011 12:05:49 PM - granted resource %MediaID%
10/21/2011 12:05:49 PM - granted resource HP.ULTRIUM5-SCSI.001
10/21/2011 12:06:54 PM - mounted; mount time: 00:01:05
10/21/2011 12:06:54 PM - Info bptm(pid=4328) EV2000
10/21/2011 12:06:54 PM - Info bptm(pid=4328) INF - Waiting for positioning of media id %MediaID% on server %MasterServer% for reading.
10/21/2011 12:06:54 PM - positioning %MediaID% to file 14
10/21/2011 12:07:44 PM - positioned %MediaID%; position time: 00:00:50
10/21/2011 12:07:44 PM - begin reading
10/21/2011 12:07:44 PM - Info bptm(pid=4328) waited for empty buffer 1 times, delayed 6 times
10/21/2011 12:07:44 PM - end reading; read time: 00:00:00
10/21/2011 12:07:44 PM - Info bptm(pid=4328) completed reading backup image
10/21/2011 12:07:44 PM - Info bptm(pid=4328) EXITING with status 0 <----------
10/21/2011 12:10:37 PM - Info bpbrm(pid=324) %Backup Proxy Host% is the host to restore data
10/21/2011 12:10:37 PM - Info bpbrm(pid=324) reading file list from client
10/21/2011 12:10:40 PM - connecting
10/21/2011 12:10:40 PM - Info bpbrm(pid=324) starting bptm
10/21/2011 12:10:43 PM - Info tar32(pid=2036) Restore started
10/21/2011 12:10:43 PM - connected; connect time: 00:00:03
10/21/2011 12:10:43 PM - Info bptm(pid=4808) start
10/21/2011 12:10:43 PM - started process bptm (4808)
10/21/2011 12:10:43 PM - Info bpdm(pid=4808) reading backup image
10/21/2011 12:10:43 PM - Info bptm(pid=4808) using 30 data buffers
10/21/2011 12:10:43 PM - Info bptm(pid=4808) spawning a child process
10/21/2011 12:10:43 PM - Info bptm(pid=4808) child pid: 6224
10/21/2011 12:10:43 PM - Info bptm(pid=6224) start
10/21/2011 12:10:43 PM - started process bptm (6224)
10/21/2011 12:10:43 PM - mounted
10/21/2011 12:10:43 PM - Info bptm(pid=4808) %MediaID%
10/21/2011 12:10:43 PM - requesting resource %MediaID%
10/21/2011 12:10:43 PM - granted resource %MediaID%
10/21/2011 12:10:43 PM - granted resource HP.ULTRIUM5-SCSI.001
10/21/2011 12:10:44 PM - Info bptm(pid=4808) INF - Waiting for positioning of media id %MediaID on server %MasterServer% for reading.
10/21/2011 12:10:44 PM - positioning %MediaID% to file 14
10/21/2011 12:12:14 PM - positioned %MediaID%; position time: 00:01:30
10/21/2011 12:12:15 PM - begin reading
10/21/2011 12:13:58 PM - Info bptm(pid=4808) waited for empty buffer 6306 times, delayed 6307 times
10/21/2011 12:13:58 PM - end reading; read time: 00:01:43
10/21/2011 12:13:58 PM - Error bptm(pid=6224) did not receive EXIT STATUS from bprd, all blocks may not have been restored
10/21/2011 12:13:58 PM - Info bptm(pid=4808) EXITING with status 175 <----------
10/21/2011 12:14:03 PM - Info tar32(pid=2036) done. status: 175: not all requested files were restored
10/21/2011 12:14:03 PM - Error bpbrm(pid=324) client restore EXIT STATUS 175: not all requested files were restored
10/21/2011 12:15:02 PM - restored image %hostname%_1319156026 - (the restore failed to recover the requested files(5)); restore time 00:09:16
10/21/2011 12:15:03 PM - end Restore; elapsed time: 00:09:21
FlashBackup-Windows policy restore error(2817)
10-21-2011 10:41 AM
Windows 7 Does this too.... Also the KB/sec counter goes crazy during restore. I can't speak for the backup as I didn't watch it.
KB/sec on backup 32000
Kb/sec on restore 110000 (impossible as I have 1GB ethernet)
10-21-2011 12:43 PM
10-22-2011 07:25 AM
FlashBackup-Windows policy restore error(2817)
1. This error appear if you are trying to restore a VM (No Agent Installed) with a Flash Windows Policy
You need to set in the BAR the Proxy Server using for backup the Virtuals Machines in Destination, also check the Virtual restore (Not a Normal Restore)
2. Check the permissions on the ESX for the backup account. I think you maibe don't have a permission set to write to the ESX
10-23-2011 03:12 PM
Tim
There is a lot to get right for vSphere API backups but when done it should work a treat!
Just double check everything - make sure the Authenticatied user for the vCenter has enough rights - the admin account is always favorite!:
http://www.symantec.com/docs/TECH128513
http://www.symantec.com/docs/TECH130493
Also check the timeouts set, especially the client connect and client read timeouts on the Media Server / VMWare backup Host as the logs hint at a lack of response at the final stage.
10-24-2011 02:19 PM
Thanks for everyone updating. I had a production server that went down today and I have just got to looking at my other work. Somehow a Socket 7 went dead... I have never seen that before but I guess it happens... One of the main reasons I'm working on getting this going is so that we can start porting production to ESX using fault tollerant VM's so this type of thing will no longer an issue. I'm sure this highlighted the issue :)
Tomorrow I will look at this more.
I will try your solutions tomorrow and will post my findings.
Thanks again,
Tim
10-25-2011 11:40 AM
Good afternoon,
My VC version is 4.1 Build 259021
My vsphere version is 4.1 Build 348481
I'm using local accounts to access the Vcetner server and that local account has administrator privlages. The backup proxy host is using the system account to login and I would say that authentication doesn't seem to be an issue because Windows XP/Windows 2003 server VM's backup just fine.
I would like to apoligze as windows 7 does in fact backup and restore correctly. I made an error in the restore.
Also note this,
Today I went to look at the backups that were supposed to run automatically. They failed because the backup proxy host was unavailable. I then looked at the ESX host and sure enough it was powered down... I thought strange!!
I then powered it on but it didn't power on due to a disk not being available. I then looked at the data stores and sure enough the disk was there. I then looked at the config and ALL the failed windows 2k8 restores created disks and attached them to the backup proxy host. I deleted those links and the backup proxy host started..
Any ideas?
10-25-2011 02:56 PM
I assume then that you are using the hotadd functionality of the vSphere backups.
Failed backups will not remove the clones - please check through the NetBackup VMWare Guide in relation to this, along with the roubleshooting gudie (all in the DOCS folder of you installation media)
Once one has failed in this way you need to clear the clones manually before another backup can be successful.
Also note the restrictions for this type of backup in relation to the host being in the same datastore etc.
Hope this helps
11-22-2011 08:02 AM
Upgrading to 7.1.0.2 did the trick. Thanks for everyone's help on this.