Project

General

Profile

Actions

Bug #11838

closed

FRR ospf6d consumes all available memory+swap after an interface event

Added by Jim Pingle 7 months ago. Updated 7 months ago.

Status:
Needs Patch
Priority:
High
Assignee:
-
Category:
FRR
Target version:
-
Start date:
04/22/2021
Due date:
% Done:

0%

Estimated time:
Plus Target Version:
Affected Version:
Affected Plus Version:
Affected Architecture:

Description

In certain cases ospf6d will consume all RAM and swap after an interface event. For me, the easiest way to reproduce it is to run ospf6d over IPsec VTI and stop/start IPsec (not restart) or save/apply on the assigned IPsec VTI interface.

After either of those events, when IPsec reconnects ospf6d will start consuming memory excessively and lead to system instability. It may eventually recover on its own, though on 21.05/2.6.0 snapshots it appears to lead to a panic (#11839).

It happens on two systems here with 4G RAM+2G swap and 8G RAM+2G swap which are otherwise not heavily loaded so it isn't necessarily tied to actually being low on RAM, but something running away in ospf6d.

Maybe nothing we can do about this since it likely must be fixed upstream, but this can stick around for tracking it so users can know we've identified it at least.

Actions #1

Updated by Jim Pingle 7 months ago

  • Description updated (diff)
Actions #2

Updated by Jim Pingle 7 months ago

I reported this upstream to FRR: https://github.com/FRRouting/frr/issues/8711

Actions

Also available in: Atom PDF