Backup of a single disk of a server seems to hang

[ict-backupng]

I have a single windows server (out of 1500) that shows an odd behavior.

I start the backup and have all 4 streams opening up (Shadow Copy Components:\, E:\, D:\ and C:\). All streams except the C:\ one complete. When i check that specific stream i see that it always stops at the same amount of KB written.

I checked the path it says it is stuck, excluded it, but nothing makes it move beyond that point. After 180 hours the backup still runs, so apparently no timeouts are being triggered.

Where can i find (logs, anywhere) what is causing this.

Extra info:

Environment:
RedHat 7, Masterserver with NetBackup 7.7 (no updates)
RedHat 7, Mediaservers (4) NetBackup 7.7 no updates
Windows clients (~1500, NetBackup 7.7 client no updates)
RedHat clients (~400, NetBackup 7.7 client no updates)
NDMP backups

[Marianne]

You need bpbkar log on the client.

If the bpbkar folder does not exist in <install-path>\veritas\netbackup\logs, please create it before the next backup.
Change Client's logging level to 3.

If possible, create a test policy for C-drive only and run manually.
This will make troubleshooting a lot easier.

[ict-backupng]

Thanks, backup is now running, will give it a few hours (total amount of data only 20GB)

[ict-backupng]

Okay, the backup has ran for 90 minutes now and only backed up 2000640 kilobyte.

The log seems to end with regular repeats of these two messages:

13:01:11.350 [960.4264] <4> tar_base::keepaliveThread: INF - sending keepalive
13:01:16.670 [960.8080] <4> tar_backup_tfi::backup_finishfile_state: INF - catalog message: Fil - 1453 24040 4308617 -1 84 33216 root;Administrators@BUILTIN root;SYSTEM@NT:AUTHORITY 1293 1477640228 1477640228 1477640228 /C/ProgramData/McAfee/Common Framework/AgentEvents/20161028093708731670600000744.xml

I cannot figure out what this means though, so i hope someone else is adapt at reading this. I attached the entire log. Server name is replaced by the text SERVER, ipaddresses have had the first two octets replaced by 999.999.

Any help more then appreciated :)

[Lowell_Palecek]

Bpbkar sends the catalog message after it has backed up a file. I see no reason for it to repeat. The implication is that it's backing up the file over and over again.

Can you exclude your McAfee data? What happens if you suspend McAfee during your backup?

I see in the log that the snapshot process (bpfis) associated with this backup succeeded. I don't see much else because it's a level 1 log. Marianne asked you for level 3. Also set the General Logging level in the Windows client subtree to 2.

[ict-backupng]

Odd, even when i check right now, global logging for this client is set to 3.

I will rerun the backup with the additional logging levels for the windows client turned to 2 as well.

[ict-backupng]

Temporary conclusion. After excluding the mcafee files it indeed runs beyond the 2GB mark from before. Odd that from 1500 servers this one seems to need this.

Will check everything next week tuesday (friday and monday national holidays here) and confirm. Will also mark the post of Lowell_Palecek as sollution then.

[Lowell_Palecek]

You are right. The global logging level was 3 on your previous backup. I missed that.

Can you exclude the McAfee folder, or suspend McAfee during the backup? Have you configured Windows Open File Backup?

<Sorry about redundancy. We were typing at the same time.>