Hi Reload,
In case the FL status vanishes from the cluster status output, it could be because of the auto-recovery process that kicks in. Please verify if your situation matches with the case mentioned in the below KB:
https://kb.juniper.net/InfoCenter/index?page=content&id=KB32508&actp=RSS
Due to the fabric link auto-recovery feature, there is no need to reboot the secondary node to restore the status; all the FPC cards of the secondary node (or the node with failures) will auto soft-restart. Due to the FPC auto soft-restart, all interfaces on secondary (or the node with failures) including the control link will temporarily go down. Both RG0 and RG1+ will change to a lost status.
This could be an expected case after some upgrades but in case you run into this issue very often, you might consider getting this investigated by JTAC as this would need analysis of logs.
If auto-recovery does not kick in, it could be that the fabric probes packets are not received by the node on which you could see FL set, it could be the JSRPD daemon stuck on the node that could not process any probe packet sent by the other node.
Hope this helps 🙂
Please mark "Accepted Solution" if this helps you solve your query.
Kudos are always appreciated!