Bug 1461655

+++ This bug was initially created as a clone of Bug #1461649 +++

Description of problem:

* start glusterd
* take statedump
* glusterd crashes

Thread 1 (Thread 0x7f535fd71700 (LWP 1557)):
#0  glusterd_dump_priv (this=<optimized out>) at glusterd-statedump.c:243
#1  0x00007f5367aaf655 in gf_proc_dump_xlator_info (top=<optimized out>) at statedump.c:502
#2  0x00007f5367aafb92 in gf_proc_dump_info (signum=signum@entry=10, ctx=0xd1e010) at statedump.c:837
#3  0x00000000004089b9 in glusterfs_sigwaiter (arg=<optimized out>) at glusterfsd.c:2083
#4  0x0000003866e07d14 in start_thread (arg=0x7f535fd71700) at pthread_create.c:309
#5  0x0000003866af168d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:115
(gdb) l
248	                                                "pmap[%d].type", port);
249	                        gf_proc_dump_write (key, "%d", pmap->ports[port].type);
250	                        gf_proc_dump_build_key (key, "glusterd",
251	                                                "pmap[%d].brickname", port);
252	                        gf_proc_dump_write (key, "%s",
253	                                            pmap->ports[port].brickname);
254	
255	                }
256	                /* Dump client details */
257	                glusterd_dump_client_details (priv);
(gdb) p pmap
$2 = (struct pmap_registry *) 0x0
(gdb) l glusterd-statedump.c:243
238	                /* Dump peer details */
239	                GLUSTERD_DUMP_PEERS (&priv->peers, uuid_list, _gf_false);
240	
241	                /* Dump pmap data structure from base port to last alloc */
242	                pmap = priv->pmap;
243	                for (port = pmap->base_port; port <= pmap->last_alloc;
244	                     port++) {
245	                        gf_proc_dump_build_key (key, "glusterd", "pmap_port");
246	                        gf_proc_dump_write (key, "%d", port);
247	                        gf_proc_dump_build_key (key, "glusterd",
(gdb) p pmap
$3 = (struct pmap_registry *) 0x0

I think since there are no bricks, pmap is NULL resulting in crash. This _MAY_ not happen when there are volumes and bricks associated with glusterd

Version-Release number of selected component (if applicable):


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info:

--- Additional comment from Red Hat Bugzilla Rules Engine on 2017-06-15 01:27:10 EDT ---

This bug is automatically being proposed for the current release of Red Hat Gluster Storage 3 under active development, by setting the release flag 'rhgs‑3.3.0' to '?'. 

If this bug should be proposed for a different release, please manually change the proposed release flag.

--- Additional comment from Atin Mukherjee on 2017-06-15 01:46:53 EDT ---

You are right. This is because the pmap object is created when glusterd allocates the the port for the first time and that will happen only when the first volume is started.