How to get the current (statistic) value from the "Max Multiplex"?

sanadmin_stadtdo · ‎2021-03-29

Hello community,

we have problems with disconnections ("unexpected network error" - often after 90 minutes) with volumes attached to Citrix servers (and only with Citrix). While searching for a solution I came across the parameter "max-mpx" for CIFS multiplex settings - for the affected SVM this is currently set to 255.

In the past, "cifs stat" could be used to determine what values were actually being used though and it looked like (Example!): "Max Multiplex = 459, Max pBlk Exhaust = 0, Max pBlk Reserve Exhaust = 0".

Where can I find this value under Ontap 9.7? When I collect values with " statistics start -object cifs ... " I don't find it there for example.

Any idea?

Many thanks and best regards

Michael

jcolonfzenpr · ‎2021-03-29

Hi:

Take a look to this KB:

https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/CIFS_Max_Multiplex_and_how_to_increase_maximum_number_of_simultaneous_...

[vserver cifs options show -vserver <vserver_name> -instance (in Clustered Data ONTAP diag mode)]

Hope this helps!

Jonathan Colón | Blog | Linkedin

sanadmin_stadtdo · ‎2021-03-29

Hello,

thanks for the message; but I was already aware of the article.

It refers to ONTAP 7 and ONTAP 8 and unfortunately does not explain how to determine under ONTAP 9 (9.7 in our case) which values "Max Multiplex" assumes in operation - "cifs stat" does not work here:

Error: show failed: Default sample not found. To collect a sample, use the "statistics start" and "statistics stop" commands. To view the data sample, use the "statistics show" command with the "-sample-id" parameter.

When using "statistic starts -object ..." I do not know in which "object" "Max Multiplex" can be found (object cifs or cifs_ctx is not).

Best regards

Michael

parisi · ‎2021-03-30

Max mpx only applies to SMB1. This KB article covers max mpx:

https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/CIFS_Max_Multiplex_and_how_to_increase_maximum_number_of_simultaneous_...

So mpx likely doesn't apply to your issue.

SMB 2 and 3 leverage credits in ONTAP.

https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/How_ONTAP_implements_SMB_crediting

I'd suggest opening a case to get to the bottom of your disconnects.

paul_stejskal · ‎2021-03-30

Agreed with Justin. However, you can do some debugging yourself. First I'd start with this KB as a launchpad: https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/How_to_troubleshoot_performance_issues_in_ONTAP_9. Once you go there, you probably will find ONTAP latency is low. Check the related articles linked to get tips to troubleshoot offbox. Basically if you think you have network issues, you probably should be grabbing a tcpdump on the appropriate interface and then viewing in Wireshark.

sanadmin_stadtdo · ‎2021-04-08

Hello,

I have to elaborate; the volume in question is on SSD, latencies are below 0.3ms and IOPs average 36. It doesn't look like there are any real performance problems here, but an availability problem of the file share - it crashes several times a day (there are no error messages in the event log of the cluster). The volume is used for Citrix; it provides user directories for LotusNotes.

A Notes consultant suspects that it is due to a limitation of "Open File Handler" or similar parameters. During my search for similar errors, I came across the parameter "Max Multiplex", which has probably led to similar errors in the Citrix environment - with the wrong specifications.

Now I'm trying to find out what value this "Max Multiplex" actually assumes in productive operation - but I don't know how.

Or do you have other ideas on how I can check the permanent availability of the share?

Many thanks

Michael

parisi · ‎2021-04-08

As mentioned, max multiplex is only an issue here if you are using SMB1 for the Citrix shares, which you most likely are not.

In the link I posted earlier, they give you a cluster command to view those stats (it's buried later in the article):

https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/CIFS_Max_Multiplex_and_how_to_increase_maximum_number_of_simultaneous_...

[statistics show-periodic -object cifs:vserver -instance vs1 interval 1 -iterations 10 -counter max_commands_outstanding|commands_outstanding (in Clustered Data ONTAP diag mode)]

This works on my system running the next release of ONTAP so it should work for yours:

tme-a300-efs01::*> statistics show-periodic -object cifs:vserver -instance DEMO -iterations 10 -counter max_commands_outstanding|commands_outstanding
tme-a300-efs01: cifs:vserver.DEMO: 4/8/2021 09:48:01
max
commands commands Complete Number of
outstanding outstanding Aggregation Constituents
----------- ----------- ----------- ------------
0 40 Yes 2
0 40 Yes 2
0 40 Yes 2
tme-a300-efs01: cifs:vserver.DEMO: 4/8/2021 09:48:08
max
commands commands Complete Number of
outstanding outstanding Aggregation Constituents
----------- ----------- ----------- ------------
Minimums:
0 40 - -
Averages for 3 samples:
0 40 - -
Maximums:
0 40 - -

But you'd probably be better served using this command to measure CIFS stats:

::> set diag

::*> statistics start -object smb1|smb1_ctx|smb2|smb2_ctx|cifs

That will collect a bunch of CIFS stats until you run "statistics stop."

I'd also suggest checking "event log show" for clues on why shares might be disconnecting.

What share properties are in use here? Is this a FlexVol or a FlexGroup?

paul_stejskal · ‎2021-04-08

https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/High_user_latency_seen_but_ONTAP_latency_is_low I'd start here for tips.