[Bugs] [Bug 1528558] /usr/sbin/ glusterfs crashing on Red Hat OpenShift Container Platform node

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1528558



--- Comment #2 from Raghavendra G <rgowdapp at redhat.com> ---
(gdb) thr 1
[Switching to thread 1 (Thread 0x7fba6ffff700 (LWP 64279))]
#0  0x00007fba76602b17 in __wb_preprocess_winds
(wb_inode=wb_inode at entry=0x7fba6434cd60) at write-behind.c:1325
1325            offset_expected = holder->stub->args.offset
(gdb) p *holder
$13 = {all = {next = 0x7fba65756520, prev = 0x7fba65756520}, todo = {next =
0x7fba643e8240, prev = 0x7fba6434cd88}, lie = {next = 0x7fba65756540, 
    prev = 0x7fba65756540}, winds = {next = 0x7fba65756550, prev =
0x7fba65756550}, unwinds = {next = 0x7fba65756560, prev = 0x7fba65756560}, wip
= {
    next = 0x7fba65756570, prev = 0x7fba65756570}, stub = 0x0, write_size =
20480, orig_size = 131072, total_size = 0, op_ret = 131072, op_errno = 0,
refcount = 0, 
  wb_inode = 0x7fba6434cd60, fop = GF_FOP_WRITE, lk_owner = {len = 8, data =
'\000' <repeats 1023 times>}, iobref = 0x0, gen = 9, fd = 0x7fba648d99d0,
wind_count = 9, 
  ordering = {size = 131072, off = 1179648, append = 0, tempted = -1, lied =
-1, fulfilled = -1, go = 0}}
(gdb) p &holder->todo
$14 = (list_head_t *) 0x7fba65756530
(gdb) p holder->lie
$15 = {next = 0x7fba65756540, prev = 0x7fba65756540}
(gdb) p &holder->lie
$16 = (list_head_t *) 0x7fba65756540
(gdb) p holder->todo
$17 = {next = 0x7fba643e8240, prev = 0x7fba6434cd88}
(gdb) p &holder->todo
$18 = (list_head_t *) 0x7fba65756530
(gdb) p holder->winds
$19 = {next = 0x7fba65756550, prev = 0x7fba65756550}
(gdb) p &holder->winds
$20 = (list_head_t *) 0x7fba65756550
(gdb) p holder->unwinds
$21 = {next = 0x7fba65756560, prev = 0x7fba65756560}
(gdb) p &holder->unwinds
$22 = (list_head_t *) 0x7fba65756560
(gdb) p holder->wip
$23 = {next = 0x7fba65756570, prev = 0x7fba65756570}
(gdb) p &holder->wip
$24 = (list_head_t *) 0x7fba65756570
(gdb) p holder->stub
$25 = (call_stub_t *) 0x0

Some observations:
* holder->stub is NULL causing the crash
* holder->refcount is 0, so the "holder" is likely to be freed
* holder->todo is not an empty list. However, holder is not part of any other
lists.

-- 
You are receiving this mail because:
You are the assignee for the bug.