used call_rpcs_sync function maybe cause crash or infinite loop

[xihong08]

Bug Report We used pegasus 2.0 and pegasus_shell disk_capacity command to query cluster disk's Usage sometimes disk_capacity -d cause crash, sometimes cause infinite loop, it Occurs occasionally. I tracked the pegasus_shell dead and cpu 100% when i frequent use disk_capacity -d the stack roughly as follows：

#0 in std::_Rb_tree_increment(std::_Rb_tree_node_base const*) 
#1 in std::_Rb_tree_const_iterator  gcc10/include/10.2.0/bits/stl_tree.h:376
#2 in std::_Rb_tree    stl_tree.h:2531
#3 in std::_Rb_tree    stl_tree.h:2542
#4 in std::map           stl_map.h:1069
#5 in call_rpcs_sync   replication_ddl_client.h:306

rpcs.erase(rpc.first) cause this proplem The above bug most likely multithreading cause by Insecure operation at variable rpcs

[acelyc111]

Thanks for your report, we'll fix it later.

[VirendraYadav1234]

I want to work on this issue please assign me this issue

[acelyc111]

I want to work on this issue please assign me this issue

@VirendraYadav1234 Thanks, assigned.