Network and Storage Protocols

Copy from local drive to CIFs share stalls


We have clients complaining for the same issue:
 copy from their local computer or VDI (Citrix) is to cifs share provided by us is so slow that it stalls, even for small files. 


The problem occurred after the network share has been migrated from another vendor (Dell) to NetApp, several users complain that the shares that are still at the other vendor are quick and the migrated ones are slow. Not only the copy stalls but sometimes MS office files on the share appear as corrupt and big pictures in them do not load properly.


Already did the following investigation and tests:

- ping and traceroute from client to data lif and vice versa are perfect and quick.

- SMB1 is disabled but clients are using Win10.

- Symantec Endpoint Protection is place on VDIs and laptops but policies are handled on company level, clients do not have rights to modify.

- AV scan is place, but has been disabled for the tests, no effect on speed
- fpolicy is off

- cifs changenotify was removed for the tests, no effect on speed

- have run tcpdump during test but they appear to be empty, like there was no incoming data

- clients can write to the share.
- not all clients experience the same.

Can I check somehow during a test with a client if they are trying to use SMB1?
Did anyone experience something alike?



> ping and traceroute from client to data lif and vice versa are perfect and quick.

Have you also tested to make sure you don't have a jumbo frame mismatch?  It's easy to miss and can be hard to find.   

Thank you, but this usually effects one in a hundred users. Also, we have MTU 9000 everywhere.