Date: Feb 2010
Affected Product & Version : NVBU 8.5; NSD 1.01
Affected Module & Version : N/A
OS Version: Any
On checking the NetVault log for a backup job writing to SmartDisk the following entries are noticed...
Transfer to the device has stalled
Transfer to the device is no longer stalled
When the Plugin on the client is unable to send network traffic to SmartDisk for a period of 15 seconds, the message "Transfer to the device has stalled" is logged.
At the point when communication is restored, the message "Transfer to the device is no longer stalled" is logged.
There is a delay sending data to SmartDisk and the cause can be network congestion or SmartDisk free space.
1 - Ensure that there are no disk space issues with the SmartDisk staging area and then
2 - check network performance. In testing to create this solution it was noticed that these messages were observed with the network running at 10Mbps. At 100Mbps and at 1G no messages were reported. The timeout can not be adjusted; 15 seconds is a long time for a network to stall.
3 - increase the block transfer size from default 32KB to at least 256KB
smartdisk.bat config --update --webdav 37453 --block-size 262144
It appears that bigger packets on certain networks help smartdisk stream data more effectively therefore reducing the number of times the device need to stall and stopping the message channel from dropping
BakBone SFDC Solution Number: 00005498