Ubuntu
ejabberd package

epam crashes instantly on ejabberd start

Bug #1767101 reported by Eugene Crosser on 2018-04-26

This bug affects 1 person

Affects		Status	Importance	Assigned to	Milestone
	ejabberd (Ubuntu)	Fix Released	Undecided	Unassigned

Bug Description

I try to switch from jabberd2 to ejabberd and I need PAM authentication. I have freshly installed `ejabberd` 18.01-2 and `erlang-p1-pam` 1.0.3-3 packages on bionic, and I enabled PAM authentication in `/etc/ejabbed/ejabberd.yml`:

```
##
## Authentication using PAM
##
auth_method: pam
pam_service: "jabber"
```

and `/etc/pam.d/jabber` is this:

```
auth sufficient pam_unix.so likeauth nullok nodelay
account sufficient pam_unix.so
```

As suggested in the debian README, I have systemctl override like this:

```
[Service]
PrivateDevices=
PrivateDevices=false
NoNewPrivileges=
NoNewPrivileges=false
```
Every time I restart jabberd service, I get a number of such entries in the crash.log:

```
2018-04-26 13:04:43 =ERROR REPORT====
** Generic server epam terminating
** Last message in was {#Port<0.19001>,{exit_status,139}}
** When Server state == {state,#Port<0.19001>}
** Reason for termination ==
** port_died
2018-04-26 13:04:43 =CRASH REPORT====
  crasher:
    initial call: epam:init/1
    pid: <0.504.0>
    registered_name: epam
    exception exit: {port_died,[{gen_server,handle_common_reply,8,[{file,"gen_server.erl"},{line,726}]},{proc_lib,init_p_do_apply,3,[{file,"proc_lib.erl"},{line,247}]}]}
    ancestors: [epam_sup,<0.486.0>]
    message_queue_len: 0
    messages: []
    links: [<0.487.0>]
    dictionary: []
    trap_exit: false
    status: running
    heap_size: 376
    stack_size: 27
    reductions: 397
  neighbours:
2018-04-26 13:04:43 =SUPERVISOR REPORT====
     Supervisor: {local,epam_sup}
     Context: child_terminated
     Reason: port_died
     Offender: [{pid,<0.504.0>},{id,epam},{mfargs,{epam,start_link,[]}},{restart_type,permanent},{shutdown,brutal_kill},{child_type,worker}]
```

followed by

```
2018-04-26 13:04:43 =SUPERVISOR REPORT====
     Supervisor: {local,epam_sup}
     Context: shutdown
     Reason: reached_max_restart_intensity
     Offender: [{pid,<0.504.0>},{id,epam},{mfargs,{epam,start_link,[]}},{restart_type,permanent},{shutdown,brutal_kill},{child_type,worker}]
```

Exit status 139 means that a child process crashed with signal 11 (SEGFAULT).
Needless to say, xmpp client authentication does not work.

Tags: