andy/glusterfs
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

geo-rep: Handle RMDIR recursively

Browse Source 
If RMDIR is recorded in brick changelog which is due to
self heal traffic then it will not have UNLINK entries for
child files. Geo-rep hangs with ENOTEMPTY error on slave.

Now geo-rep recursively deletes the dir if it gets ENOTEMPTY.

BUG: 1129702
Change-Id: Iacfe6a05d4b3a72b68c3be7fd19f10af0b38bcd1
Signed-off-by: Aravinda VK <avishwan@redhat.com>
Reviewed-on: http://review.gluster.org/8477
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Venky Shankar <vshankar@redhat.com>
Tested-by: Venky Shankar <vshankar@redhat.com>
This commit is contained in:
Aravinda VK 2014-08-12 18:19:30 +05:30 committed by Venky Shankar
parent 718f10e0d6
commit 2510af1674
1 changed files with 12 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
geo-replication/syncdaemon/resource.py
View File

@ -26,6 +26,7 @@ import subprocess
from errno import EEXIST, ENOENT, ENODATA, ENOTDIR, ELOOP
from errno import EISDIR, ENOTEMPTY, ESTALE, EINVAL
from select import error as SelectError
import shutil
from gconf import gconf
import repce
@ -616,6 +617,17 @@ class Server(object):
while True:
er = entry_purge(entry, gfid)
if isinstance(er, int):
if er == ENOTEMPTY and op == 'RMDIR':
er1 = errno_wrap(shutil.rmtree,
[os.path.join(pg, bname)],
[ENOENT])
if not isinstance(er1, int):
logging.info("Removed %s/%s recursively" %
(pg, bname))
break
logging.warn("Failed to remove %s => %s/%s. %s" %
(gfid, pg, bname, os.strerror(er)))
time.sleep(1)
else:
break