10-27-2016 12:25 AM
Hello ,
so we searching after upgrading to 7.7.3 a Solution we has open Tickets against support but now for more than 1 Month and nothing helps ...
The Setup
Inventory
Netbackup 7.7.3 on Redhat Linux 6.
Netapp 4 Node Cluster.
svm on the cluster for cifs/nfs.
I500 Tape LiB with Lto6
data(svm)
cluster(cluster svm)
nbu1 (master)
nbu2-7(media)
Config
NDMP Credentials stored for :
cluster
data
SLP:
cdot Snapshot and than BackupFromSnapshot to Tape
Policie :
BackupNDMP
Included the data svm as NDMP Node and the Volume that we want save.
Problem:
The backup Run's, after the Tape Backup is finished we get an error :
10/27/2016 08:58:27 - Info bptm (pid=10930) EXITING with status 23 <----------
10/27/2016 08:58:27 - Critical bpbrm (pid=10922) unexpected termination of client data
I we look in the the bptm log :
08:55:49.385 [10930] <2> write_data_tir: absolute block position prior to writing backup header(s) is 33813, copy 2
08:55:49.385 [10930] <2> write_data_tir: block position check: actual 33813, expected 33813
08:55:49.385 [10930] <2> io_write_back_header: drive index 1, np-hms09_1477551074, file num = 4, mpx_headers = 0, copy 2
08:55:49.388 [10930] <2> io_write_block: ndmp_tape_write_func returned 1024
08:55:49.392 [10930] <8> retry_getaddrinfo_for_real: [vnet_addrinfo.c:1108] getaddrinfo() failed RV=-2 NAME=cluster SVC=0 errno=2
08:55:49.392 [10930] <8> retry_getaddrinfo: [vnet_addrinfo.c:948] retry_getaddrinfo_for_real failed RV=-2 NAME=cluster SVC=0
08:55:49.392 [10930] <8> vnet_cached_getaddrinfo_and_update: [vnet_addrinfo.c:1758] retry_getaddrinfo() failed RV=-2 NAME=cluster SVC=NULL
08:55:49.393 [10930] <8> vnet_cached_getaddrinfo: [vnet_addrinfo.c:1386] vnet_cached_getaddrinfo_and_update() failed 6 0x6
08:55:49.393 [10930] <2> NdmpMoverSession[0]: CreateServer: vnet_cached_getaddrinfo failed, host = cluster, status = 6, error = -2
08:55:49.393 [10930] <2> NdmpMoverSession[0]: MoverConnect method failed for IPv6. Reinitializing and attempting to use IPv4.
08:58:20.262 [10930] <2> NdmpMoverSession[0]: ndmp_mover_connect failed, status = NDMP_CONNECT_ERR
08:58:20.264 [10930] <2> NdmpMoverSession[0]: MoverConnect method failed. Reinitializing and attempting to use MoverListen method.
08:58:20.276 [10930] <2> NdmpMoverSession[0]: ndmp_mover_listen failed with address type 0x20510004, status = NDMP_CONN_TYPE_UNCONFIGURED_ERR
08:58:20.276 [10930] <16> NdmpMoverSession[0]: ERROR Start failed
08:58:20.276 [10930] <16> check_and_process_mover_tasks: NDMP Mover Client Setup failed
So we found out that we can disable the NDMP Mover than the backup run once.
After that the backup run once, the next backup get the error (these error come also if the mover is enabled.) :
10/27/2016 09:13:54 - Info bpbrm (pid=13467) Starting delete snapshot processing
10/27/2016 09:13:54 - Info bpfis (pid=13476) Backup started
10/27/2016 09:13:54 - Warning bpbrm (pid=13467) from client defr2elvbng01: Stream BackupNDMP_data+d10412+1 pid 10798 is not active.
10/27/2016 09:13:54 - Warning bpbrm (pid=13467) from client defr2elvbng01: Stream BackupNDMP_data+dd10412+1 pid 11375 is not active.
10/27/2016 09:13:54 - Warning bpbrm (pid=13467) from client defr2elvbng01: Stream BackupNDMP_data+dd10412+1 pid 12282 is not active.
10/27/2016 09:13:54 - Warning bpbrm (pid=13467) from client defr2elvbng01: Stream BackupNDMP_data+dd10412+1 pid 12975 is not active.
Operation Status: 1567
10/27/2016 09:13:55 - Info bpfis (pid=13476) done. status: 0
10/27/2016 09:13:55 - end SnapDupe Mount:Unmount Snapshot; elapsed time 0:00:02
10/27/2016 09:13:55 - Info bpfis (pid=13476) done. status: 0: the requested operation was successfully completed
10/27/2016 09:13:55 - end writing
Operation Status: 0
10/27/2016 09:13:55 - begin SnapDupe Mount: End Notify Script
10/27/2016 09:13:55 - Info RUNCMD (pid=13488) started
10/27/2016 09:13:55 - Info RUNCMD (pid=13488) exiting with status: 0
Operation Status: 0
10/27/2016 09:13:55 - end SnapDupe Mount: End Notify Script; elapsed time 0:00:00
Operation Status: 1567
Only one NDMP backup of a snapshot per backup ID is allowed (1567)
So now im lost.
I has to change the names so it's posible that nothing changed alll...
So any help would be very welcome .
rg
marcel
Solved! Go to Solution.
11-11-2016 04:18 AM
Thanks all for the posts.
So i found out after DebugLevel 9 that the Cluster-node that served the data build up active an NDMP Sesssion with the Master Server.
These Connection was frobidden in the Firewall. So for all reading this, it is really nessary that all node managment lifs can comminicate with your master / media server.
It is not sufficiant that only the cluster lif can communicate with the master.
rg
marcel
10-27-2016 01:28 AM
I do not have 1st-hand experience - just bookmarks for the day I may need to.
Please compare your config with these TNs:
Clustered Data ONTAP (cDot) 8.3 configuration outline for use with NetBackup 7.7 Cluster Aware Backup (CAB) backup:
http://www.veritas.com/docs/000025335
This one seems to be for pre-7.7:
Additional NetApp ONTAP 8.2 C-Mode NDMP configuration notes for NetBackup :
http://www.veritas.com/docs/000021393
10-27-2016 02:29 AM
Been awhile since I have worked with NetApp, but the first thing I notice is this line:
<8> retry_getaddrinfo_for_real: [vnet_addrinfo.c:1108] getaddrinfo() failed RV=-2 NAME=cluster SVC=0 errno=2
which seem to indicate a name resolution problem.
Another thing that used to be an issue was that the tape drive order on the filer changed unless you used the Netapp version of persistent binding. Was some entries in the rc file back then.
10-27-2016 08:22 AM
NetBackup status code: 1567
Message: Only one NDMP backup of a snapshot per backup ID is allowed
Explanation: A backup policy of type NDMP refers to an SLP that contains more
than one Backup From Snapshot operation.
Recommended Action: Do one of the following:
¦ Modify the SLP so that it contains only one Backup From Snapshot operation.
¦ Modify the backup policy so that it uses a different SLP.
11-11-2016 04:18 AM
Thanks all for the posts.
So i found out after DebugLevel 9 that the Cluster-node that served the data build up active an NDMP Sesssion with the Master Server.
These Connection was frobidden in the Firewall. So for all reading this, it is really nessary that all node managment lifs can comminicate with your master / media server.
It is not sufficiant that only the cluster lif can communicate with the master.
rg
marcel