Bug 1759829 - write-behind xlator generate coredump,when run "glfs_vol_set_IO_ERR.t"
Summary: write-behind xlator generate coredump,when run "glfs_vol_set_IO_ERR.t"
Keywords:
Status: CLOSED UPSTREAM
Alias: None
Product: GlusterFS
Classification: Community
Component: write-behind
Version: 6
Hardware: x86_64
OS: Linux
unspecified
high
Target Milestone: ---
Assignee: bugs@gluster.org
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2019-10-09 08:38 UTC by Rongjunfei
Modified: 2020-03-12 12:15 UTC (History)
2 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2020-03-12 12:15:53 UTC
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Embargoed:

Attachments (Terms of Use)

Description Rongjunfei 2019-10-09 08:38:33 UTC 
Description of problem:
run cmd below core dump
"prove -vf ./tests/bugs/gfapi/glfs_vol_set_IO_ERR.t"

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


How reproducible:


Steps to Reproduce:
1.modify write-behind source code,
  change option "trickling-writes" default value to "off",
  so write requests can aggregate 
2.compile source code
  ./configure --enable-debug
  make -j12
  make install
3.prove -vf ./tests/bugs/gfapi/glfs_vol_set_IO_ERR.t


Actual results:
glfs_vol_set_IO_ERR: write-behind.c:721: wb_inode_destroy: Assertion `list_empty(&wb_inode->todo)' failed.
./tests/bugs/gfapi/../../include.rc: line 342: 25924 Aborted                 (core dumped) ./tests/bugs/gfapi/glfs_vol_set_IO_ERR centos patchy /var/log/glusterfs/glfs_vol_set_IO_ERR.log

Expected results:


Additional info:
(gdb) bt
#0  0x00007f833b81a2c7 in raise () from /usr/lib64/libc.so.6
#1  0x00007f833b81b9b8 in abort () from /usr/lib64/libc.so.6
#2  0x00007f833b8130e6 in __assert_fail_base () from /usr/lib64/libc.so.6
#3  0x00007f833b813192 in __assert_fail () from /usr/lib64/libc.so.6
#4  0x00007f83260378c1 in wb_inode_destroy (wb_inode=0x7f82f8000b90) at write-behind.c:721
#5  0x00007f8326037931 in wb_forget (this=<optimized out>, inode=<optimized out>) at write-behind.c:2823
#6  0x00007f833aea737e in __inode_ctx_free (inode=inode@entry=0x1fe1e68) at inode.c:411
#7  0x00007f833aea8ab2 in __inode_destroy (inode=0x1fe1e68) at inode.c:432
#8  inode_table_prune (table=table@entry=0x7f83100593c0) at inode.c:1681
#9  0x00007f833aeaa7cb in inode_table_destroy (inode_table=0x7f83100593c0) at inode.c:1972
#10 0x00007f833aeaa8a1 in inode_table_destroy_all (ctx=ctx@entry=0x1f18410) at inode.c:1891
#11 0x00007f833bbbb5de in pub_glfs_fini (fs=0x1f18270) at glfs.c:1354
#12 0x000000000040103c in main (argc=4, argv=0x7fffb7e3fd88) at ./tests/bugs/gfapi/glfs_vol_set_IO_ERR.c:157
Comment 1 Rongjunfei 2019-10-09 09:14:19 UTC 
1.write-behind.c
wb_inode_destroy 
{
   ......
   GF_ASSERT(list_empty(&wb_inode->todo));
   ......
}

2.glfs_vol_set_IO_ERR.c
write_something(glfs_t *fs);
{
   ......
   glfs_write(fd, buf, WRITE_SIZE, 0);
   ......
}

this function will left some write requests in write-behind
than "glfs_fini" will  call "wb_inode_destroy"
so generate coredump
Comment 2 Worker Ant 2020-03-12 12:15:53 UTC 
This bug is moved to https://github.com/gluster/glusterfs/issues/854, and will be tracked there from now on. Visit GitHub issues URL for further details
Note You need to log in before you can comment on or make changes to this bug.