Hello people,
Our ndmp filers can connect to the backup server but backupping the data is not going to complete.
The others clients such as Windows file servers and linux machines backup state is perfect.The backup media master server is NB 8.1.2 on Windows Server 2016 fully patched.
The backup server can connect to the NDMP host, runs for 10 - 20 minutes while writing data and then stops.
Can someone give usefull information?
Sound like this tech note:
STATUS CODE 42: NetBackup NDMP backups fail while backing up large number of files with the error message, Error ndmpagent (pid=8998) connection 0x081871d8 ndmp_message_process_one failed, status = 18 (NDMP_XDR_DECODE_ERR)
https://www.veritas.com/support/en_US/article.100017530
But I would also take a look from the Netapp side;
Hello Nicolai,
Thank you the reply and suggestion to the tech note.
I already have added a MAX_ENTRIES_PER_ADD file to D:\veritas\netbackup with the max value 100000
There are no strange logs at netapp side with debugging enabled related to ndmp netbackup.
This is more than likley a difference in "tuning" values, between what you had in the old Master, and what you have since the upgrade..
One thing that may help to prove this is to do a backup with "set hist=n" as the first line of the includes/selections list, this will treat the backup as a single large dump, and not "index" every file in it, thus proving that the data can be read / written and that its the data being sent regarding the individual files that is the problem.
https://www.veritas.com/support/en_US/doc/24427212-127304472-0/id-SF940177686-127304472
NOTE with the "hist=n" setting you will NOT be able to individual file restores, you have to recover the entire volume, but it will help rule out the need to tune things...
Looking at OLD questions regarding a similar issue https://vox.veritas.com/t5/NetBackup/NDMP-XDR-DECODE-ERR/td-p/442148
this could also be due to JUMBO frames.
Regards
Simon
Thank you Simon for your reply and ideas regarding this situation. I would like to be able to recover single files so set hist=n is not an option. The ndmp backups a netapp with esx server and if 1 server needs to be restored then the entire is being restored. I can not find the jumbo frames on windows server 2016 > nic > properties > configure > advanced.
**** Post edited to remove unrelated URL ****
Please verify if there are any defunct process already running for the same job on the media servers?? If so, please kill those processes and start the jobs. You may consider to restart the NetBackup on media server where these NDMP jobs are running and kill all the netbackup process before you start the NetBackup on the media server. I have exprienced a similar scenario.
Tape_Archived, thanks your your reply.
We had our monthly patch tuesday so I updated all server in our network. Restarted the ndmp server and netbackup master/media server. There were no jobs left. 1st new job was ndmp backup and same result.
----
