Ceph » rgw

In a multisite environment, below crash is observed when restarting radosgw servers -

(gdb) bt
#0 0x00007ffff751fa41 in RGWRESTConn::RGWRESTConn (this=0x555556c40000,
cct=<optimized out>, store=0x555556aeeb00,
_remote_id="da140565-8396-4939-befd-25bd553a646e", remote_endpoints=...,
_api_name=std::optional<std::string> = {...}, _host_style=PathStyle)
at ../src/rgw/rgw_rest_conn.cc:25
#1 0x00007ffff77eff91 in (anonymous namespace)::make_peer_connections<std::map<std::_cxx11::basic_string<char>, RGWZoneGroup> > (zonegroups=..., store=0x555556aeeb00)
at /usr/include/c++/10/bits/basic_string.h:907
#2 (anonymous namespace)::MasterTrimEnv::MasterTrimEnv (this=<optimized out>,
dpp=<optimized out>, store=0x555556aeeb00, http=<optimized out>,
num_shards=<optimized out>) at ../src/rgw/rgw_trim_mdlog.cc:231
#3 0x00007ffff77f5cd6 in MetaMasterTrimPollCR::MetaMasterTrimPollCR (interval=...,
num_shards=64, http=0x555556bb68b8, store=0x555556aeeb00, dpp=0x555556bb67c8,
this=0x555555ea0800) at ../src/rgw/rgw_trim_mdlog.cc:660
#4 create_meta_log_trim_cr (dpp=dpp@entry=0x555556bb67c8, store=0x555556aeeb00,
http=http@entry=0x555556bb68b8, num_shards=<optimized out>, interval=...)
at ../src/rgw/rgw_trim_mdlog.cc:723
#5 0x00007ffff7503e68 in RGWSyncLogTrimThread::process (this=0x555556bb6780,
dpp=0x555555e92be8) at ../src/common/config_proxy.h:120
#6 0x00007ffff74bc558 in RGWRadosThread::Worker::entry (this=0x555555e92bb0)
at ../src/rgw/rgw_rados.cc:355
#7 0x00007ffff572a432 in start_thread () from /lib64/libpthread.so.0
#8 0x00007ffff52709d3 in clone () from /lib64/libc.so.6
(gdb) p (rgw::sal::RadosStore*)store
$3 = (rgw::sal::RadosStore *) 0x555556aeeb00
(gdb) p *$
$4 = {<rgw::sal::Store> = {
_vptr.Store = 0x7ffff7f6e870 <vtable for rgw::sal::RadosStore+16>},
rados = 0x555555dd6400, user_ctl = 0x0, luarocks_path = "",
zone = std::unique_ptr<rgw::sal::RadosZone> = {get() = 0x0}}
(gdb)

>>> This seems to be regression with change - https://github.com/ceph/ceph/pull/45623/ .