As part of regular operations yesterday I was draining traffic off cr2-esams, so we could reset "pic 0" to reconfigure the speeds of the QSFP ports in the device after a change.
As part of the normal process we use for this I issued the below command to enable the BGP "graceful shutdown" feature, which lowers local-preference for all BGP learnt routes on the box, and adds the well-known "gshut" community to any it is announcing to signal to peers not to use it.
set protocols bgp graceful-shutdown sender
For whatever reason a minute or so later (did not appear to be right away) rpd (routing-protocol daemon, the service that controls routing protocols on JunOS) crashed:
Jun 15 15:00:10 cr2-esams kernel: pid 20071 (rpd), jid 0, uid 0: exited on signal 6 (core dumped)
Opening JTAC case with Juniper to see if they can explain this and if it is a known bug. Something similar may have occurred last week on cr2-drmrs on the same JunOS version, however there things recovered by themselves and quicker. But there were some unexplained ping losses to the device during that maintenance I didn't expect, and the same command had been issued.