Create a Post
cancel
Showing results for 
Search instead for 
Did you mean: 
gemechisd
Contributor

Checkpoint gateway Memory Usage

We have checkpoint security gateway 7000 appliance. Recently the memory usage is above 90%. How can we know, which process is running more memory?

 

 

0 Kudos
35 Replies
CheckPointerXL
Advisor
Advisor

free -h command?

 
0 Kudos
gemechisd
Contributor

Do I have to be on Expert mode to run free -h?

0 Kudos
G_W_Albrecht
Legend Legend
Legend

Look at the prodedure for memory here: https://supportcenter.checkpoint.com/supportcenter/portal?eventSubmit_doGoviewsolutiondetails=&solut...

CCSE / CCTE / CTPS / CCME / CCSM Elite / SMB Specialist
0 Kudos
Chris_Atkinson
Employee Employee
Employee

Also what do you see with the output from the "top" command?

Which version & Jumbo is the gateway installed with?

Note the 7000 series appliances can be upgraded to 64GB RAM where required. But first we should determine if the memory is consumed by a high number of connections or a memory leak etc?

CCSM R77/R80/ELITE
0 Kudos
gemechisd
Contributor

The output of top command is attached below

0 Kudos
svori
Collaborator
Collaborator

Please post output of top and sorted by memory usage.

That is SHIFT+m when top is running

0 Kudos
gemechisd
Contributor

Here it is

0 Kudos
the_rock
Legend
Legend

@gemechisd 

Just to be 100% sure its not any debugs running in the background, can you issue below 2 commands, wait few mins and see if anything changes?

fw ctl debug 0

fw ctl debug -x

Cheers,

Andy

0 Kudos
gemechisd
Contributor

What effect does this commands have?

0 Kudos
the_rock
Legend
Legend

It disables and turns off any debugs on the firewall that might be currently running.

0 Kudos
gemechisd
Contributor

Okay. Does it have any effect on gateway at the moment if I run the commands you have listed?

 

0 Kudos
the_rock
Legend
Legend

100% does not have any effect. I ran it at least 200 times and never had an issue.

0 Kudos
the_rock
Legend
Legend

To check debugs that could be on, you can refer to below and what Tim Hall suggested. 

Andy

 

https://community.checkpoint.com/t5/General-Topics/How-to-check-debug-command/td-p/32373

0 Kudos
gemechisd
Contributor

So, what could the possible reason that can make the memory to run on 98%?

0 Kudos
the_rock
Legend
Legend

There can be many reasons. I would say, from my experience, its usually specific process thats causing it, or it simply could be firewall that does not have enough physical RAM, so has to use swap. It also could be amount of traffic, so thats why running those commands I sent you would give us pretty good idea.

You can also run -> tw ctl multik print_heavy_conn and it would show you connections that are heavily "utilized"

Hope that helps.

Andy

0 Kudos
the_rock
Legend
Legend

Also, to add to my last response, cpstat if really good command. If you run it, it will give you all the options. Idea is thia essentially -> run cpstat, then flag on the left, then -f then whatever word (flag_ on the right of the table

few examples:

cpstat fw -f all

cpstat fw -f policy

cpstat vpn -f all

cpstat blades -f fw

Hope those help as well.

Andy

0 Kudos
PhoneBoy
Admin
Admin

I see a ton of dlpu processes here.
What blades are enabled on the gateway?
Execute the command enabled_blades in expert mode.

0 Kudos
gemechisd
Contributor

Below is the attached screenshot of enabled blades.

0 Kudos
Chris_Atkinson
Employee Employee
Employee

Please note if Web Extraction is enabled it will use additional RAM per sk145773.

Do you have any snmp monitoring of your concurrent connections & ram/memory consumption over time available?

Otherwise TAC can assist with a procedure to perform further analysis and eliminate memory leaks and similar so a determination on upgrading memory can be made.

CCSM R77/R80/ELITE
0 Kudos
gemechisd
Contributor

@PhoneBoy @Chris_Atkinson @the_rock 

Here is the enabled blades feature output.

[Expert@]# enabled_blades
fw vpn cvpn urlf av appi ips identityServer anti_bot ThreatEmulation mon Scrub
0 Kudos
the_rock
Legend
Legend

K, so you have both threat emulation AND threat extraction, wondering if that could be related to this. Can you run cpview again quick when you log in and see what is the current usage at the moment?

Lets do remote when you have 15-20 mins, I can have a look.

Let me know.

Cheers mate.

0 Kudos
Chris_Atkinson
Employee Employee
Employee

How much traffic, connections & throughput is the appliance handling whilst these blades are enabled?

Drip feeding the answers is not helping. Without all the information we cannot suggest anything different than we already have.

CCSM R77/R80/ELITE
0 Kudos
the_rock
Legend
Legend

I think, in all fiarness, @gemechisd is trying to give us as much info, but its bit tricky without possible remote session...just my 2 cents.

0 Kudos
Chris_Atkinson
Employee Employee
Employee

Perhaps but likely TAC would have all the data in a cpinfo already if they've been engaged... 

CCSM R77/R80/ELITE
the_rock
Legend
Legend

Good point Chris. @gemechisd do you have TAC case open? If you have cpinfo, Im happy to review it myself as well if you can share it securely.

Cheers mate.

0 Kudos
the_rock
Legend
Legend

I would check below commands:

free -m

top

ps -auxw

cpview

cpwd_admin list

0 Kudos
Sorin_Gogean
Advisor

Hello @gemechisd ,

 

So you have an Quantum 7K with (I assume as you didn't specify) 32Gb of RAM and with all the blades enabled (as you showed in the last post, and you complain that the memory is 90-95% utilized ?

I mean, what would you expect? Isn't that what's the memory there for? to be utilized ?

In our case we're having OLD 15600 in our clusters (currently at 32Gb of RAM), and with almost all the blades enabled (we're not doing VPN from the appliances, therefore IPsec VPN and Mobile Access are not active in our case, and also Content Awareness, but we have also the DLP, QOS, Identity Awareness and HTTPS Inspection that are eating their chunk of Memory 😊) we're around 75 - 90% of memory. 
Therefore, this weekend we're adding another 32Gb so we will be at 64Gb tops, so we're expecting the appliance to be around 50-70% of memory utilization, as we also intend to make use of SandBlast Blades (Threat Emulation and Extraction) by adopting Autonomous Threat Prevention. 
So please do "Execute the command enabled_blades in expert mode" like @PhoneBoy asked, and not a SmarConsole screenshot.... (we have HTTPS Inspection enabled and that is not showing on License Status page!!!!!) 


Bottom-line, since you enabled all that was possible, and your appliance isn't MAX out from resource perspective, I would say it's normal behavior.  If you want to lower that, then disable blades that are not used, and configure/fine-tune the others and you will be at an 75% maybe. But my recommendation, would be to go ahead and MAX the memory from the appliance, and then you won't bother about that part. Plus you will be able to make use of other blades, without any resource concerns.

 

Thank you,

PS: that reminds me of some ticket/alert of riverbeds equipment's in our environment, being all the time at 90-95% of the memory. and my response on that, was "is there any issue with the data Optimization that RB does, if not, why we care of smth like that since the RB functions in it's parameters and there were not thresholds triggered, like the RB guys have set/defined...." The box was normal to be on 90-95% memory utilization as it's made to process data in a specific way.
PS2: it's a bit tough to say what program/blade uses what quantity of memory, since they split in multiple instances that are running in parallel, therefore the recommendation to dump the unused blades is the first one, followed by the ticking of the remaining ones.
PS3: in our case, we are getting currently to an 98% of memory utilization, when we spike 300-500K connections (in the past we got also 1MIL connections) and that was increasing memory consumption, therefore our decision to MAX our the resources, in this case memory.

0 Kudos
gemechisd
Contributor

@Sorin_Gogean 

Yeah we have 7K appliance with 32GB RAM. 

But the memory utilization is increasing before 2 weeks. And after that we were not able to see any logs on our SMS. I will do the enabled_blades command and see if there is anything new/I'll share it. 

0 Kudos
Sorin_Gogean
Advisor

So, you're saying that you see memory increase in last 2 weeks, or after an appliance restart, as soon as we put traffic through it, in 2 weeks time we're seeing an increase of memory ?

Also in regards to the missing logs, is this used as a Management also or you have a separate management/log server or appliance ? 

As for the blades, can you tell what are you actually using, or you are unaware as you just enable all possible ones and didn't configure half of them...

 

Ty,
PS: I really can't understand why you're not presenting all the things from the beginning, in order for us to know your environment, and point you properly 🙄.

Leaderboard

Epsum factorial non deposit quid pro quo hic escorol.

Upcoming Events

    CheckMates Events