Restarting httpsd without knowing PID?

[dan]

Just looking through the 6.4.8 Known Issues and found this:

721487

FortiGate often enters conserve mode due to high memory usage by httpsd process.

This seems to be similar to the WAD issue:

712584

WAD memory leak causes device to go into conserve mode.

I have a (sad) workaround for the WAD issue by regularily issuing diagnose test application wad 99 with auto-script.

However, there is no similar way for the httpsd.

It looks like I'd need to ssh into the box and first identify the httpsd PID with diagnose system top and afterwards issue diag sys kill 11 <process-Id>

• Is there a better, automated way?
• When there is no active Web-GUI session, is httpsd restarted?
• In your experience, when does the httpsd high memory usage comes up?

Conserve mode scares me off, since I already had twice to drive to client locations to get an FG-60 and an FG-61 working again.

Thanks

Daniel

[Toshi_Esumi]

Fortunately I didn't have to do this for quite some time but "fnsysctl killall" still seems to work according to this post.
https://github.com/bluecmd/fortigate_exporter/issues/62

Toshi

[ede_pfau]

Sadly, one FG-200E has had exactly the same memory leak issue running on v6.2.9 a couple of weeks ago.

What I did (from advice from TAC supporter) was write a script with "fnsysctl killall httpsd" and create an automation stitch which was triggered by the even "enter conserve mode". Worked like a charm - the instant the log message was written, the script ran and reduce memory usage by 40%.

(The ultimate fix was to upgrade to v6.4.8. Good that I didn't know about the "Known issue" you cited.)

config system automation-trigger
   edit "enter_conserve_mode"
      set event-type low-memory
   next
end

config system automation-action
   edit "release_memory"
      set action-type cli-script
      set required enable
      set script "fnsysctl killall httpsd"
      set accprofile "super_admin"
   next
   edit "auto_high_memory_email"
      set action-type email
      set email-to "monitor@mysite.de"
      set email-subject "myFGT: stitch alert: conserve mode"
   next
end

config system automation-stitch
   edit "auto_high_memory"
      set trigger "enter_conserve_mode"
      set action "release_memory" "auto_high_memory_email"
   next
end

[dan]

Thanks to @ede_pfau  and @Toshi_Esumi . I will try fnsysctl killall httpsd.

It did not show any effect while manually testing, but I believe I will implement the given script. 

Better sure than sorry..

Dan

[pavankr5 · Accepted solution]

Hello @dan,

Please check this article on Technical-Tip-How-to-restart-kill-all-processes
https://community.fortinet.com/t5/FortiGate/Technical-Tip-How-to-restart-kill-all-processes-with-fns...
let us know if this helps 

Thanks,

Pavan