cluster/dht: Clear clean_dst flag on target change

If the target of a file migration was changed because
of min-free-disk limits, the dst_fd was closed but the
clean_dst flag was not set to false. If the file could
not be created on the new target for some reason, the
ftruncate call to clean up the dst was sent on the now
invalid fd causing the process to deadlock.

Change-Id: I5bfa80f519b04567413d84229cf62d143c6e2f04
BUG: 1469029
Signed-off-by: N Balachandran <nbalacha@redhat.com>
Reviewed-on: https://review.gluster.org/17735
Smoke: Gluster Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.org>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Raghavendra G <rgowdapp@redhat.com>
This commit is contained in:
N Balachandran 2017-07-10 15:45:04 +05:30 committed by Raghavendra G
parent f7a450c17f
commit bd71ca4fdf

View File

@ -1599,10 +1599,14 @@ dht_migrate_file (xlator_t *this, loc_t *loc, xlator_t *from, xlator_t *to,
}
syncop_close (dst_fd);
dst_fd = NULL;
old_target = to;
to = new_target;
clean_dst = _gf_false;
/* if the file migration is successful to this new target, then
* update the xattr on the old destination to point the new
* destination. We need to do update this only post migration
@ -1617,6 +1621,7 @@ dht_migrate_file (xlator_t *this, loc_t *loc, xlator_t *from, xlator_t *to,
} else {
gf_msg (this->name, GF_LOG_INFO, 0, 0, "destination for file "
"- %s is changed to - %s", loc->path, to->name);
clean_dst = _gf_true;
}
}