Here is the backtrace info in core file
(gdb) bt
#0 0x7f79da065b7f in raise () from /lib64/libpthread.so.0
#1 0x7f79e5303563 in reraise_fatal (signum=6) at
/usr/src/debug/ceph-16.2.10-0.el8.x86_64/src/global/signal_handler.cc:332
#2 handle_fatal_signal (signum=6) at
Hi- it settled back to having 4 MDS services, and the file system is up and
running. However the 4 MDS services are just on 2 hosts:
[root@story ~]# ceph fs status home
home - 6 clients
RANK STATE MDSACTIVITY DNSINOS DIRS CAPS
0active home.hiho.mssdyh
Hi,
your subject is "...two monitors per host" but I guess you're
asking for MDS daemons per host. ;-) What's the output of 'ceph orch
ls mds --export'? You're using 3 active MDS daemons, maybe you set
"count_per_host: 2" to have enough standby daemons? I don't think an
upgrade would