NDMP Full back up keep failling

Hello @Egor Skepko 

It looks like the job is running for some time before the failure, though performance is very poor. The Data Socket (data receive connection from the filer) is coming into the MediaAgent at ~1.5Gb\hr. Once Commvault has the data, we process it around 30Gb\hr, still slow but the receive connection is much slower.

After some time the network drops between the Filer and the MediaAgent and then further down we can see the MediaAgent failing to connect to processes through its own loopback address.

Its possible something like Anti Virus, low system resources, or port exhaustion is causing this issue. There may also be multiple contributing factors, the slow network can cause threads to timeout and cause similar behavior, but we woudln’t expect to see failures connecting to the loopback.

Event Viewer, ProcMon, & Wireshark may be of assistance here.

I am having an issue responding with the log cuts in the response so I’ve put them in a notepad document.

Thank you,

Collin
 

@Collin Harper Yea i sow that performance are really bad/slow. The issue is that alote of small files need to be scan on HNAS and that take to long to do it, then the job failed because taking to long. 

Hello @Egor Skepko 

The job shouldn’t fail because its taking too long unless you have a rule in place to kill jobs running longer than [x] amount of time.

If you have a lot of small files, it might be worth considering and testing with the Windows Filesystem agent for Network Shares. You can back up all the NAS shares with all the functionality and benefits of the Filesystem Agent.

Getting Started with Network Share for CIFS Shares - https://documentation.commvault.com/11.26/expert/92615_getting_started_with_network_share_for_cifs_shares.html

Thank you,
Collin

@Collin Harper Yes we have like 20M smalls files of data, but NDMP not faster then CIFS?