[BETA2] Snapshot overflow

[MyKey0815]

Hello,

I want to Backup a server based on UCS (Whezzy) to a remote share. The installation is working fine. The job can create successfull.

When the job is running 4-5 Minutes the job is "failed".

Code: Select all

     2016-09-05 09:31:15 Job started at 2016-09-05 09:31:15
     2016-09-05 09:31:15 Starting backup
     2016-09-05 09:31:15 Creating volume snapshot                                                   00:00:01
     2016-09-05 09:31:19 Backing up BIOS bootloader on /dev/sda                                     00:00:02
     2016-09-05 09:31:21 Backed up /dev/sda 489.1MB at 55.4MB/s                                     00:00:09
     2016-09-05 09:33:21 [error] Backing up vg_ucs 1.7GB at 70.2MB/s (1%)                           00:01:45
     2016-09-05 09:35:14 [error] Snapshot overflow.
     2016-09-05 09:35:14 [error] Failed to upload disk.
     2016-09-05 09:35:14 [error] Agent failed to process method {DataTransfer.SyncDisk}.
     2016-09-05 09:35:14 [error] Exception from server: Snapshot overflow.
     2016-09-05 09:35:14 [error] Unable to retrieve next block transmission command. Number of al...
     2016-09-05 09:35:14 [error] Failed to download disk.
     2016-09-05 09:35:14 [error] Failed to perform backup

[PTide]

Hi,

What kind of a remote share do you use, is it SMB or NFS? Could you please try to run the backup one more time and see of the error repeats? Please double-check that no disconnection of the share occurs during the backup.

Thanks

[MyKey0815]

I try it agin. Same error again, sorry.

Can you tell me, how can i check the connection/disconnection of the share? Do you know a tool that i can make a stresstest? I dont have a idea?

Thanks for help

[PTide]

Can you tell me, how can i check the connection/disconnection of the share?

Try to copy any file to the share right after the backup has failed. If the share is disconnected then copy operation should fail.

Do you know a tool that i can make a stresstest?

Try to copy /usr/src directory to the shared folder and see if it works.

Also please share an archived /var/log/veeam directory so we can review logs.

Thank you

[MyKey0815]

Thanks for your reply.

i have test it again. Direct after the error i try to copy some files to the share. It works fine. All files can be copied to to share.

Now i have the whole /var/log/veem Folder attached in this link
https://we.tl/IK3PrmDgen

[PTide]

So, what's the type of the share that you use? Is it Windows share?

[MyKey0815]

No, it´s a Samba4 (4.1.13-Ubuntu) Share with following definition:

Code: Select all

[veeam]
   comment = Veeam Datensicherungs Freigabe
   path = /HDD/SDA252_4000G/Veeam
   browseable = yes
   read only = no
   guest ok = yes
   writeable = yes
   force directory mode = 2770
   create mask = 0644
   directory mask = 2777

[PTide]

Do any files appear on the share during the backup?

[MyKey0815]

Do any files appear on the share during the backup?

Yes. The directory and a file (named: Backup_Full_2016-09-08T040016.vbk) is created. The file is 960.470 KB size

[PTide]

Do you see any .vbk files on the UCS server?

[MyKey0815]

In which directory i must looking for this file? Do you mean with "UCS Server" the Source Server that should be backuped?

[PTide]

Do you mean with "UCS Server" the Source Server that should be backuped?

Yes.

In which directory i must looking for this file?

Somewhere under /tmp, there should be "veeam" directory inside.

My assumption is that Samba got disconnected in the middle of the proccess therefore the session continued to transfer data right to the same device where vg_ucs is located causing the snapshot to overflow. Another possible reason for overflow is a very high data change rate, that can be resolved if you set up a higher snapshot size limit.



Thanks

[MyKey0815]

I have start a new try.

When the job is failed I found in a file in /tmp/veeam/192.168.3.252veeam/ucs003090 Backup_Full.

Code: Select all

960484 -rw-r--r-- 1 nobody nogroup 983528448 Sep  9  2016 Backup_Full_2016-09-09T114705.vbk

[PTide]

Ok, so that piece of backup tells me that for some reason veeam has not managed to write to samba share. Let's check if that happened due to a lack of permission, could you please change permissions so anyone would be able to write to the share and modify files that are already there, and start backup afterwards?

Thanks

[MyKey0815]

They can write to the samba share. The first 960MB will written to the share - they are still on the network.

The permissions is set to Everyone

[PTide]

They can write to the samba share. The first 960MB will written to the share - they are still on the network.

Could you please provide me with info about backup directory and files permissions on samba share? What I need to know is ls -la /HDD/SDA252_4000G/Veeam/ucs003090 Backup_Full

When the job is failed I found in a file in /tmp/veeam/192.168.3.252veeam/ucs003090 Backup_Full.

Did you find a .vbk with the same name (Backup_Full_2016-09-09T114705.vbk) on the share?

Thanks

[MyKey0815]

Here the result

Code: Select all

root@backuppc:~# ls -la /HDD/SDA252_4000G/Veeam/ucs003090\ Backup_Full/
total 960492
drwxrwxr-x 2 nobody nogroup      4096 Sep  9 11:47 .
drwxrwxrwx 3    777 root         4096 Sep  9 11:47 ..
-rw-r--r-- 1 nobody nogroup 983528448 Sep  9 11:51 Backup_Full_2016-09-09T114705.vbk