[EX] Commit operation causes LACPD_TIMEOUT

Summary:

It may be seen that a commit operation on EX4550/EX4200 Series switches is causing a spike in CPU utilization followed by LACP or BFD timeout. However, this issue has been resolved in Junos OS Releases 12.3R4-S3 and 12.3R5. This article gives some details.

Symptoms:

After performing commit action on EX4550/EX4200 Series switches, the Link Aggregation Control Protocol (LACP) is seen to time out as shown below:

Sep 15 18:16:23  Switch chassisd[5185]: CHASSISD_HWDB_ERROR: hwdb call hwdb_picinfo_inst_iter_start failed, pic 61646, error No such attribute
Sep 15 18:16:23  Switch last message repeated 3 times
Sep 15 18:16:25  Switch /kernel: KERN_LACP_INTF_STATE_CHANGE: lacp_update_state_userspace: cifd xe-4/0/30 - ATTACHED state - acting as standby link  <<<
Sep 15 18:16:27  Switch /kernel: KERN_LACP_INTF_STATE_CHANGE: lacp_update_state_userspace: cifd xe-4/0/30 - CD state - ready to carry traffic
Sep 15 17:25:51  Switch /kernel: KERN_LACP_INTF_STATE_CHANGE: lacp_update_state_userspace: cifd xe-4/0/30 - ATTACHED state - acting as standby link <<<
Sep 15 17:25:51  Switch /kernel: KERN_LACP_INTF_STATE_CHANGE: lacp_update_state_userspace: cifd xe-4/0/30 - CD state - ready to carry traffic
Sep 15 17:25:51  Switch lacpd[5159]: LACPD_TIMEOUT: xe-4/0/30: lacp current while timer expired current Receive State: CURRENT

Cause:

On EX4550/EX4200 Series switches running Junos OS Release 12.2R5 or Release 12.3R3 onward, commit operations might cause a spike in CPU utilization, resulting in timeout of time sensitive applications/protocols such as LACP and Bidirectional Forwarding Detection (BFD).

Solution:

This issue has been resolved in Junos OS Release 12.3R4-S3 and Release 12.3R5 as indicated by PR898097.

Knowledge Base