Bug 1917382

Core corresponds to sssd-common-2.4.0-5.el8

$ file coredump 
coredump: ELF 64-bit LSB core file, 64-bit PowerPC or cisco 7500, version 1 (SYSV), SVR4-style, from '/usr/libexec/sssd/sssd_be --domain implicit_files --uid 0 --gid 0 --logger=file', real uid: 0, effective uid: 0, real gid: 0, effective gid: 0, execfn: '/usr/libexec/sssd/sssd_be', platform: 'power8'

#0  0x000000013a22bc90 in dp_client_handshake_timeout (ev=<optimized out>, te=0x10032a63100, t=..., ptr=<optimized out>)
    at src/providers/data_provider/dp_client.c:153
#1  0x00007fffb1ad06a4 in tevent_common_invoke_timer_handler () from /lib64/libtevent.so.0
#2  0x00007fffb1ad0908 in tevent_common_loop_timer_delay () from /lib64/libtevent.so.0
#3  0x00007fffb1ad2234 in epoll_event_loop_once () from /lib64/libtevent.so.0
#4  0x00007fffb1acf6e0 in std_event_loop_once () from /lib64/libtevent.so.0
#5  0x00007fffb1ac7cd8 in _tevent_loop_once () from /lib64/libtevent.so.0
#6  0x00007fffb1ac8084 in tevent_common_loop_wait () from /lib64/libtevent.so.0
#7  0x00007fffb1acf610 in std_event_loop_wait () from /lib64/libtevent.so.0
#8  0x00007fffb1ac8178 in _tevent_loop_wait () from /lib64/libtevent.so.0
#9  0x00007fffb2559fec in server_loop (main_ctx=0x10032a2f810) at src/util/server.c:730
#10 0x000000013a2196f4 in main (argc=<optimized out>, argv=0x7fffe7ce1528) at src/providers/data_provider_be.c:772


(gdb) frame 0
#0  0x000000013a22bc90 in dp_client_handshake_timeout (ev=<optimized out>, te=0x10032a63100, t=..., ptr=<optimized out>)
    at src/providers/data_provider/dp_client.c:153
153	    if (name != NULL && strcmp(name, be_name) == 0) {


(gdb) p be_name
$1 = 0x0


(gdb) list
148	    be_name = dp_cli->provider->be_ctx->sbus_name;
149	
150	    talloc_set_destructor(dp_cli, NULL);
151	
152	    name = sbus_connection_get_name(dp_cli->conn);
153	    if (name != NULL && strcmp(name, be_name) == 0) {
154	        /* This is the data provider connection. Just free the client record
155	         * but keep the connection opened. */
156	        talloc_zfree(dp_cli);
157	        return;


(gdb) p dp_cli->provider->be_ctx->sbus_name
$2 = 0x0


(gdb) p *(dp_cli->provider->be_ctx)
$6 = {ev = 0x10032a2f4f0, cdb = 0x10032a309d0, domain = 0x10032a55b20, identity = 0x10032a3daf0 "%BE_implicit_files", 
  conf_path = 0x10032a421e0 "config/domain/implicit_files", sbus_name = 0x0, uid = 0, gid = 0, override_space = 0 '\000', sr_conf = {
    scope = SESSION_RECORDING_SCOPE_NONE, users = 0x0, groups = 0x0, exclude_users = 0x0, exclude_groups = 0x0}, be_fo = 0x10032a42270, 
  be_res = 0x10032a40cd0, online_cb_list = 0x0, run_online_cb = false, offline_cb_list = 0x0, run_offline_cb = false, reconnect_cb_list = 0x0, 
  unconditional_online_cb_list = 0x0, offline = false, check_if_online_ptask = 0x0, mon_conn = 0x0, refresh_ctx = 0x0, check_online_ref_count = 0, 
  check_online_retry_delay = 0, provider = 0x0, last_dp_state = -1}



Hi Pavel, 

is it expected `sbus_name` can be NULL here?

In theory, yes. Because we currently set be_ctx->sbus_name after data provider is initialized in dp_initialized(), there may be some async tasks that happens before we get there. But the timeout is set to 5 seconds so something unexpected must have happened that we did not get in dp_initialized() for so long. Logs would be nice to understand it. But to mitigate it we can set it in or before dp_init_send().

Pushed PR: https://github.com/SSSD/sssd/pull/5467

* `master`
    * 0c6924b8d474daf35ee30d74e5496957e503b206 - SBUS: set sbus_name before dp_init_send()

*** Bug 1930540 has been marked as a duplicate of this bug. ***

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (sssd bug fix and enhancement update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2021:1666