Large NDMP backups end with a progress timeout when NetBackup 5220 Appliance is used as the Media Server.
|Article:TECH190563|||||Created: 2012-06-07|||||Updated: 2012-09-14|||||Article URL http://www.symantec.com/docs/TECH190563|
NDMP backups would fail with a Status 99. Reviewing the logs ndmpagent and ndmp logs, it is apparent that NetBackup is waiting for information to be received from the filer. The filer logs indicate that it is waiting for a confirmation from NetBackup that it can continue sending data. After the configured NDMP_PROGRESS_TIMEOUT value has gone by, the backup fails with a Status 99. In the examples we saw the MSDP storage pool was used for these backup attempts.
<2> bpbrm Exit: client backup EXIT STATUS 99: NDMP backup failure
In the ndmpagent log search for ReadFromSocket function to see when was the last time data was received from the filer.
05/23/2012 08:42:25.503 [Debug] NB 51216 ndmpagent 134 PID:8082 TID:46961296557984 File ID:134 [No context] 6 [NdmpMoverSideShm::ReadFromSocket] bytesLeft 262144 bytesRead 18808
NetBackup 5220 Appliance v. 2.0.2
There was a bug fixed in NetBackup 5220 Appliance version 2.0.2 where some of the network parameters in the /etc/sysctl were not set. Some of the settings of the network parameters introduced in version 2.0.2 were causing the problem.
The /etc/sysctl.conf file on the Appliance should be edited so that the following lines are set as shown here:
net.core.rmem_max = 33554432
net.core.wmem_max = 33554432
net.core.rmem_default = 262144
net.core.wmem_default = 262144
net.ipv4.tcp_rmem = 32768 262144 33554432
net.ipv4.tcp_wmem = 32768 262144 33554432
The rest of the parameters should be commented out -
# vm.dirty_background_ratio = 5
# vm.dirty_ratio = 20
# vm.swappiness = 20
# net.ipv4.tcp_window_scaling = 0
# net.ipv4.conf.all.send_redirects = 0
# net.ipv4.ipfrag_low_thresh = 524288
# net.ipv4.ipfrag_high_thresh = 1048576
# net.ipv4.tcp_tw_reuse = 1
# net.ipv4.tcp_tw_recycle = 1
# net.ipv4.tcp_max_syn_backlog = 4096
# net.ipv4.tcp_syncookies = 0
# net.ipv4.tcp_sack = 0
# net.ipv4.tcp_dsack = 0
# net.ipv4.tcp_keepalive_time = 510
# net.ipv4.tcp_keepalive_probes = 3
# net.ipv4.tcp_keepalive_intvl = 3
Large NDMP backups ends with Status 99 going to MSDP pool; 5220
After upgrade to Appliance version 2.0.2 large NDMP backups are failing with status 99.
Article URL http://www.symantec.com/docs/TECH190563