h_partial_key_nelem

[Arskaz]

Hi!

Any ideas, what do these messages indicate at messages?

kernel: [fw4_2];h_partial_key_nelem: called for table id 399 with no index table!

Not created a case yet, but those are appearing sometimes, when also some traffic problems are seen.

BR, Arskaz

[the_rock]

I found 2 articles on support site when searching that error, but they both indicate its fixed in certain jumbo version, does not reallyt say whats causing it.

[Arskaz]

OK. I couldn't find that at all from support site 😉

Maybe expert -level is not enough...

So I assume, that it is fixed at the latest R80.30 GA JHF. Have not seen that messages with it.

[the_rock]

I think it should be enough, try below : - )

https://supportcenter.checkpoint.com/supportcenter/portal?eventSubmit_doGoviewsolutiondetails=&solut...

https://supportcenter.checkpoint.com/supportcenter/portal?eventSubmit_doGoviewsolutiondetails=&solut...

If you google your error, its first 2 links that come up.

Andy

[G_W_Albrecht]

As long as your traffic problems have gone away also, the issue is fixed 😎. The cited SKs are not for [fw4_2];h_partial_key_nelem but other kernel messages - one from a GW freeze, one from a crash...

[the_rock]

Thats true, articles are not, I was just stating thats what came up when I googled it : - )

[Tal_Paz-Fridman]

Hi

Can you share exactly what version you are running (including JHF).

Thanks

Tal

[Arskaz]

R80.30Take: 236

[Tal_Paz-Fridman]

Thank you from the information.

I am taking it with the relevant R&D owners.

[Duane_Toler]

Hey,

Just wanted to add my vote to this thread.  I am helping a customer with various odd issues on their R80.30 JHF 236 HA cluster.  I had this same/similar message appear in fwk.elg repeatedly (every minute or two):

[8 Oct  9:37:00][fw4_0];[vs_0];h_partial_key_nelem: called for table id 466 with no index table!

[8 Oct  9:37:00][fw4_5];[vs_0];h_partial_key_nelem: called for table id 466 with no index table!

[8 Oct  9:37:56][fw4_0];[vs_0];h_partial_key_nelem: called for table id 466 with no index table!

[8 Oct  9:37:56][fw4_5];[vs_0];h_partial_key_nelem: called for table id 466 with no index table!

[8 Oct  9:39:01][fw4_0];[vs_0];h_partial_key_nelem: called for table id 466 with no index table!

[8 Oct  9:39:01][fw4_3];[vs_0];h_partial_key_nelem: called for table id 466 with no index table!

[8 Oct  9:39:18][fw4_0];[vs_0];h_partial_key_nelem: called for table id 466 with no index table!

[8 Oct  9:39:18][fw4_5];[vs_0];h_partial_key_nelem: called for table id 466 with no index table!

[8 Oct  9:41:04][fw4_0];[vs_0];h_partial_key_nelem: called for table id 466 with no index table!

[8 Oct  9:41:04][fw4_1];[vs_0];h_partial_key_nelem: called for table id 466 with no index table!

# fw tab -t 466 -s

HOST                  NAME                                ID #VALS #PEAK #SLINKS

localhost             466                                466   952  1039       0

I didn't find anything in SecureKnowledge (Expert level, too), but this thread did come up (obviously).  

The info below is just FYI, for how I came across the above error, to explain my environment a little more:

For this customer, I'm suspecting they have network and VPN performance issues with combination of SNDs overloaded and CoreXL being crushed due to VPN + Medium Path (QoS and IPS enabled).  They have Multi-Queue NICs, but we need to re-arrange their connections to take full advantage; this is a pending task.

I have an EVAL for them right now, with 6 FWK cores configured, but some cores are hitting 99% (with others at 83% at the same time).  Since it's R80.30, I notice SNDs are no longer limited as before, so we have as many SNDs as CPU cores.  This gateway just needs more Multi-Queue. 🙂 R80.40  is in our plan, too. 

I'm also looking at sk165853 (cpwd_q_init_ke), but I haven't done a debug yet to confirm this is related.  It's in my plan, as well.

Thanks, and I hope this helps!

[Tal_Paz-Fridman]

Hi

According to R&D these messages are harmless.

We will issue an SK as well.

Tal

[Rick_Dimmick]

I am seeing these same messages in /var/log/messages, full of them!  I was unable to find anything in SK but found this in Checkmates.  Was an SK ever created?  How do we get rid of these logs filling up the messages file?