cancel
Showing results for 
Search instead for 
Did you mean: 

track directory

gkman
Level 5

we have two servers (windows) that are running out of space.

after a quick check we found that most space is consumed by the growing size of the track directory of the Netbackup client.

I googled if it is OK to empty the folder- but all I found is a guide to move its location. (https://www.veritas.com/support/en_US/article.HOWTO77409)

we don't really have another location for the track logs, and I don't mind leaving them where they are if I can empty them every now and then when they get too big (right now they are about 10-20gb)

can anyone please explain what these logs are and if it is OK to delete them every so often? I know that they have to do with accelerator, but are they just used for troubleshooting if needed, or are they used for processing the next backup or similar?

thanks in advance for the help.

1 ACCEPTED SOLUTION

Accepted Solutions

sdo
Moderator
Moderator
Partner    VIP    Certified

1) Sounds like a track log for a cluster hasn't been configured, and so a cluster failover will cause the backup of the cluster resources to use a stale track log.  Some links for track log placement:

How to redirect the NetBackup Accelerator track log to a different location
http://www.veritas.com/docs/000068640

...and a very useful post from @MortenSeeberg here, which seems to be more informative/useful than the tech note just above:

https://vox.veritas.com/t5/Articles/Accelerator-backup-of-Windows-Cluster-systems/ta-p/813666

.

2) I not aware of exactly how Accelerator works, but I did once look inside some of the files and they are some kind of binary proprietary format for which I never seen any documentation, so I'm not sure that it's going to be possible to explain exactly why the track log is twice the size of the other.

3) Have a look at the folder name structure, look at where the policy names fit in the structure, and see if you have some track logs for old policy names that no longer exist, as these can be deleted, e.g. from a time before the client was clustered.

View solution in original post

6 REPLIES 6

sdo
Moderator
Moderator
Partner    VIP    Certified

The track logs are integral to NetBackup Accelerator.  If you delete them then at best Accelerator will have more work to do during the next full backup, or at worst may break Accelerator.

There is some argument for not backing-up the /track folder tree, but IMO there really is no argument for deleteing the entrire /track folder tree unless you are ceasing to use Accelerator.

If you have moved a client from policy to another, then you should be able to delete the old no longer used policy names from within the track folder.

sdo
Moderator
Moderator
Partner    VIP    Certified

You didn't state your NetBackup version, nor whether the two servers in question are NetBackup Client or NetBackup Media Server or NetBackup Master Server?

And, whether you are performing NetBackup Accelerator based backups of any NetApp filers?

netbackup is 7.7.2 across all servers and clients. 
and no netapp filers are being backed-up (how is it relevant?)

I am still unclear about the track logs- I understand they are integral for the accelerator, but what exactly is saved in them? why are they so big? it seems (compared to other servers) like an anomaly that needs at least some cleaning- I have a hard time believing it requires more than 10GB.

sdo
Moderator
Moderator
Partner    VIP    Certified

1) How many files make up a full backup of the server?

2) There is a formula to estimate the size:

https://vox.veritas.com/t5/NetBackup/Accelerator-Track-Log-Size-This-a-valid-formula/td-p/739851

...what size do you get when you use this formula?

now that I see the formula I guess it is possible for it to reach 10GB.

however, I did not mention before (I thought it was irrelevant) that both servers are part of the same failover cluster. raising some new questions:

1) shouldn't the track log sit on a cluster resource that will be failed over too? what happens when the servers are in two policies, one for the cluster and the second for the OS- where should the track log be?
2) why is the track log 10GB on one server and twice as much on the second?

thanks for the help

sdo
Moderator
Moderator
Partner    VIP    Certified

1) Sounds like a track log for a cluster hasn't been configured, and so a cluster failover will cause the backup of the cluster resources to use a stale track log.  Some links for track log placement:

How to redirect the NetBackup Accelerator track log to a different location
http://www.veritas.com/docs/000068640

...and a very useful post from @MortenSeeberg here, which seems to be more informative/useful than the tech note just above:

https://vox.veritas.com/t5/Articles/Accelerator-backup-of-Windows-Cluster-systems/ta-p/813666

.

2) I not aware of exactly how Accelerator works, but I did once look inside some of the files and they are some kind of binary proprietary format for which I never seen any documentation, so I'm not sure that it's going to be possible to explain exactly why the track log is twice the size of the other.

3) Have a look at the folder name structure, look at where the policy names fit in the structure, and see if you have some track logs for old policy names that no longer exist, as these can be deleted, e.g. from a time before the client was clustered.