Juniper EX – LLDP Crash

I ran into an issue with LLDP on a virtual chassis of EX4200 switches. When enabling LLDP and LLDP-MED, the LLDP process would constantly crash and restart. Each time it crashes it was creating a core dump. In the messages log file I could see these errors:

Apr 22 12:44:37  router1 init: lldpd-service (PID 69355) terminated by signal number 11. Core dumped!
Apr 22 12:44:37  router1 init: lldpd-service is thrashing, not restarted
Apr 22 12:44:46  router1 mgd[69485]: UI_OPEN_TIMEOUT: Timeout connecting to peer 'lldpd-service'
Apr 22 12:45:21  router1 dumpd: Core and context for lldpd saved in /var/tmp/lldpd.core-tarball.1.tgz
Apr 22 12:45:30  router1 dumpd: tar: lldpd.core.1.gz: Cannot stat: No such file or directory tar: Error exit delayed from previous errors
Apr 22 12:45:30  router1 dumpd: Unable to create core tarball /var/tmp/lldpd.core-tarball.1.tgz
Apr 22 12:45:41  router1 dumpd: tar: lldpd.core.1.gz: Cannot stat: No such file or directory tar: Error exit delayed from previous errors
Apr 22 12:45:41  router1 dumpd: Unable to create core tarball /var/tmp/lldpd.core-tarball.1.tgz
Apr 22 12:45:43  router1 mgd[69231]: UI_CHILD_SIGNALED: Child received signal: PID 69232, signal Broken pipe: 13, command='/usr/libexec/ui/show-support'

The configuration I had was nothing special:

{master:0}[edit]
[email protected]# show protocols lldp
interface all;

{master:0}[edit]
[email protected]# show protocols lldp-med
interface all;

As a work around, deleting lldp-med fixes the issue, after that no more crashing of LLDP.

This is most likely a bug with the JunOS release running on these switches, 12.3R4.6. It is an old release which I have not had a chance to upgrade yet, other virtual chassis on newer releases do not experience the same issue.

Posted in Juniper.

Leave a Reply

Your email address will not be published.