cancel
Showing results for 
Search instead for 
Did you mean: 

SLP for Tape out at Remote site tries to connect to client

sjvz
Level 3

We have AIR set up to replicate all of the images to a Remote Master Appliance.  Because of longer retention, we have a duplications job to tape that runs after the Replication is done.  The duplicate job is erroring out with status code 191.

2/4/2016 6:22:54 AM - requesting resource LCM_Remotemaster-hcart2-robot-tld-0
2/4/2016 6:22:54 AM - granted resource LCM_Remotemaster-hcart2-robot-tld-0
2/4/2016 6:22:55 AM - started process RUNCMD (245679)
2/4/2016 6:22:56 AM - ended process 0 (245679)
2/4/2016 6:22:57 AM - begin Duplicate
2/4/2016 6:23:13 AM - requesting resource Remotemaster-hcart2-robot-tld-0
2/4/2016 6:23:13 AM - requesting resource @aaaag
2/4/2016 6:23:13 AM - reserving resource @aaaag
2/4/2016 6:23:13 AM - reserved resource @aaaag
2/4/2016 6:23:13 AM - granted resource 100064
2/4/2016 6:23:13 AM - granted resource HP.ULTRIUM5-SCSI.003
2/4/2016 6:23:13 AM - granted resource Remotemaster-hcart2-robot-tld-0
2/4/2016 6:23:13 AM - granted resource MediaID=@aaaag;DiskVolume=PureDiskVolume;DiskPool=dp_disk_Remotemaster;Path=PureDiskVolume;StorageServer=Remotemaster;MediaServer=Remotemaster
2/4/2016 6:27:23 AM - Error bpduplicate(pid=245679) db_IMAGE() failed: cannot connect on socket (25)      
2/4/2016 6:27:23 AM - Error bpduplicate(pid=245679) Status = no images were successfully processed.      
2/4/2016 6:27:24 AM - Error bpduplicate(pid=245679) Duplicate of backupid exchange_1451963083 failed, cannot connect on socket (25).   
2/4/2016 6:27:24 AM - Error bpduplicate(pid=245679) Status = no images were successfully processed.      
2/4/2016 6:27:24 AM - end Duplicate; elapsed time: 0:04:27
no images were successfully processed (191)

 

 

based on the logs it looks like its trying to the Client.  We have opened firewall rules between the remote master and the source master and media servers , but not the client.

Is there any reason why the remote master would be trying to connect to the clients. The firewall logs are full of drops from the remote master trying to communicate back to the clients

I do have a ticket opened , but havent had much progress yet.

any thoughts

 

1 ACCEPTED SOLUTION

Accepted Solutions

mnolan
Level 6
Employee Accredited Certified

This is because it is exchange GRT. Use this:

Configuring the Exchange granular proxy host

View solution in original post

10 REPLIES 10

sdo
Moderator
Moderator
Partner    VIP    Certified

Was the original backup performed using a backup policy with GRT enabled for MS Exchange ?

mnolan
Level 6
Employee Accredited Certified

This is because it is exchange GRT. Use this:

Configuring the Exchange granular proxy host

sjvz
Level 3

mnolan , sdo

thanks much !  

i am trying to see how to stop the existing SLP and rerun it so that it picks up the change.

sdo
Moderator
Moderator
Partner    VIP    Certified

You may need to modify the previous version of the SLP, which cannot be done in the GUI.

See:

Simple guidelines on how to modify attributes of SLP version

https://www.veritas.com/support/en_US/article.TECH145081

.

...and:

How To Modify A Specific Version Of A Specific Storage Lifecycle Policy

http://www.veritas.com/docs/000014718

sjvz
Level 3

well  there was nothing to modify in the slp ,  i just dont know how to force it to look up the exchange granular backup host for the pending images , so i opened the firewall rules to the client and now i am getting database error

2/5/2016 2:52:54 PM - Error bpduplicate(pid=151367) db_IMAGE() failed: database system error (220)       
2/5/2016 2:52:54 PM - Error bpduplicate(pid=151367) Status = no images were successfully processed.      
2/5/2016 2:52:54 PM - Error bpduplicate(pid=151367) Duplicate of backupid dag01_1451952115 failed, database system error (220).    
2/5/2016 2:52:55 PM - Error bpduplicate(pid=151367) Status = no images were successfully processed.      
 

sjvz
Level 3

I have also disabled message level cataloging for GRT exchange backups under the host properties for the master. I am still getting the same error.

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified
When I look at the TN, it seems to me that SLP duplicate needs the "-granular_proxy" option. Since the TN does not mention such an option, I can only assume that the option is not available for SLPs. This should be considered as a shortcoming in NBU. Your case engineer and BL need to refer this to Engineering. What is the status of your Support case? Has it been escalated to BL?

sjvz
Level 3

Marianne

I am still waiting for the Support engineer to call me on this. He has only sent me a couple of emails so far.  He is the one who suggested disabling the message level cataloging to help with this issue. I will ask him to refer this to engineering , if he calls me.  I am not sure what BL stands for. 

I have been looking through the firewall logs and i have seen lots of dropped packets where the remote master is trying to reach the local clients.  The SLP's in those cases complete successfully ( inspite of the dropped packets ) , so i am not sure of whats causing that traffic.  its just the images associated with exchange thats failing.

The remote master is on two separate networks  - one is a Gig network and is used for management, the other network is a 10Gb backup network behind a firewall. 

Here is the case number - Veritas NetBackup Case # 21139668  , in case you have the ability to look that up.

Thanks

 

 

 

 

 

 

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

BL stands for Back Line Support.

The Engineer needs to escalate to BL if he/she cannot find a solution, who then will investigate and escalate to Engineering if there is no obvious solution.

I find it strange that we don't see this issue more often in the forums as replication followed by duplication to tape is fairly common practice - with all sorts of policy types.

You need to escalate this via your MyVeritas 'Manage case' page or by phoning Support and request escalation.
Do not wait for the engineer to phone you - he might be under the impression that he is waiting for feedback from you....

 

sjvz
Level 3

unchecked this option on the remote master  and that solved the problem . I was doing everything on the source master earlier  ....need more coffee , maybe whiskey

Capture.JPG