ONTAP Hardware

FAS3250 connection limits

AndreyOTX
2,437 Views

I am looking for the following for this device

* maximum # of concurrent/connections, and how to monitor

* maximum # of NFS/mount requests per second, and how to monitor

 

We are observing degradation. Seems to be related to mounting, as time to mount goes from milliseconds to hundreds and beyond. Application does hundreds mounts per second via UDP.

1 ACCEPTED SOLUTION

Ontapforrum
2,402 Views

I haven't come across any limitation per se protocol, ultimately it all boils down to how the data points (LIFs) are doing and their load and client's kernel capability .  What ontap version are you running  on FAS3250?

 

To start with, follow standard troubleshooting steps: ( I am assuming cdot)
1) Begin with finding an overloaded LIF by comparing the number of connections on each LIF. Look for the 'count' on each Node. May be work re-distributing LIFs between Nodes.
::> network connections active show-lifs
::> network connections active show-clients -node <node>

 

2) What NFS versions being served?

 

3) Also, if you OCUM/AIUM, just have a look around how the Nodes are doing in terms of processing on different components such as node/aggregate/volume/data-processing etc.

 

4) From client side : What logs we got, and what errors we see in the /var/log. What is the Client Linux Kernel ? Could it be port-exhaustion ?
https://www.suse.com/support/kb/doc/?id=000017508

 

This will give you some basic idea around the connections/latency overall and then you may be able to narrow down the cause. There are lots of tracing available in cdot that could help but I wouldn't go that far. If nothing helps, just raise a ticket with NetApp.

 

 

Just for reference, this may not apply to your case:

High NFS Latency with multiple connections to node LIFs from common client TCP socket
https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Systems/FAS_Systems/High_NFS_Latency_with_multiple_connections_to_node_LIFs_from_common_...

 

https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/NFSv4.1_client_Read%2F%2FWrite_operation_may_be_slow_or_hang_when_pNFS...

 

View solution in original post

2 REPLIES 2

Ontapforrum
2,403 Views

I haven't come across any limitation per se protocol, ultimately it all boils down to how the data points (LIFs) are doing and their load and client's kernel capability .  What ontap version are you running  on FAS3250?

 

To start with, follow standard troubleshooting steps: ( I am assuming cdot)
1) Begin with finding an overloaded LIF by comparing the number of connections on each LIF. Look for the 'count' on each Node. May be work re-distributing LIFs between Nodes.
::> network connections active show-lifs
::> network connections active show-clients -node <node>

 

2) What NFS versions being served?

 

3) Also, if you OCUM/AIUM, just have a look around how the Nodes are doing in terms of processing on different components such as node/aggregate/volume/data-processing etc.

 

4) From client side : What logs we got, and what errors we see in the /var/log. What is the Client Linux Kernel ? Could it be port-exhaustion ?
https://www.suse.com/support/kb/doc/?id=000017508

 

This will give you some basic idea around the connections/latency overall and then you may be able to narrow down the cause. There are lots of tracing available in cdot that could help but I wouldn't go that far. If nothing helps, just raise a ticket with NetApp.

 

 

Just for reference, this may not apply to your case:

High NFS Latency with multiple connections to node LIFs from common client TCP socket
https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Systems/FAS_Systems/High_NFS_Latency_with_multiple_connections_to_node_LIFs_from_common_...

 

https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/NFSv4.1_client_Read%2F%2FWrite_operation_may_be_slow_or_hang_when_pNFS...

 

SpindleNinja
2,349 Views

hwu.netapp.com has a few mentioned.   "Maximum number of connections - NAS" etc. 

Public