diff options
| author | Peter Portante <peter.portante@redhat.com> | 2013-01-12 22:39:24 -0500 |
|---|---|---|
| committer | Anand Avati <avati@redhat.com> | 2013-02-04 08:44:19 -0800 |
| commit | 189eaecf61091145389142b961a1c9f190299f60 (patch) | |
| tree | 5930224fc39ab58c51c476a4fc42c7c47b1ab8da /ufo | |
| parent | 2a46c8769bc2b6ad491a305ea1d38023d0e22617 (diff) | |
| download | glusterfs-189eaecf61091145389142b961a1c9f190299f60.tar.gz glusterfs-189eaecf61091145389142b961a1c9f190299f60.tar.xz glusterfs-189eaecf61091145389142b961a1c9f190299f60.zip | |
object-storage: yield during directory tree walks
Have the co-routine (greenlet) yield during long running directory
tree walks to avoid starving out other greenlets that might be
available to run.
A directory walk involves statting every file in the tree to get it
size. For large numbers of files, this can get expensive. So we yield
after every stat call, and after processing every directory.
Change-Id: I07f1dfeef6a09b5817e0c237ecc748c491d52a31
BUG: 894674
Signed-off-by: Peter Portante <peter.portante@redhat.com>
Reviewed-on: http://review.gluster.org/4380
Reviewed-by: Kaleb KEITHLEY <kkeithle@redhat.com>
Reviewed-by: Jeff Darcy <jdarcy@redhat.com>
Reviewed-by: Mohammed Junaid <junaid@redhat.com>
Tested-by: Mohammed Junaid <junaid@redhat.com>
Diffstat (limited to 'ufo')
| -rw-r--r-- | ufo/gluster/swift/common/utils.py | 25 |
1 files changed, 15 insertions, 10 deletions
diff --git a/ufo/gluster/swift/common/utils.py b/ufo/gluster/swift/common/utils.py index d35abe5825..795ddfa9d1 100644 --- a/ufo/gluster/swift/common/utils.py +++ b/ufo/gluster/swift/common/utils.py @@ -19,6 +19,7 @@ import errno import xattr import random from hashlib import md5 +from eventlet import sleep import cPickle as pickle from ConfigParser import ConfigParser, NoSectionError, NoOptionError from swift.common.utils import normalize_timestamp, TRUE_VALUES @@ -233,27 +234,30 @@ def _update_list(path, cont_path, src_list, reg_file=True, object_count=0, # strip the prefix off, also stripping the leading and trailing slashes obj_path = path.replace(cont_path, '').strip(os.path.sep) - for i in src_list: + for obj_name in src_list: if obj_path: - obj_list.append(os.path.join(obj_path, i)) + obj_list.append(os.path.join(obj_path, obj_name)) else: - obj_list.append(i) + obj_list.append(obj_name) object_count += 1 if reg_file: - bytes_used += os.path.getsize(path + '/' + i) + bytes_used += os.path.getsize(os.path.join(path, obj_name)) + sleep() return object_count, bytes_used def update_list(path, cont_path, dirs=[], files=[], object_count=0, bytes_used=0, obj_list=[]): - object_count, bytes_used = _update_list(path, cont_path, files, True, - object_count, bytes_used, - obj_list) - object_count, bytes_used = _update_list(path, cont_path, dirs, False, - object_count, bytes_used, - obj_list) + if files: + object_count, bytes_used = _update_list(path, cont_path, files, True, + object_count, bytes_used, + obj_list) + if dirs: + object_count, bytes_used = _update_list(path, cont_path, dirs, False, + object_count, bytes_used, + obj_list) return object_count, bytes_used @@ -281,6 +285,7 @@ def _get_container_details_from_fs(cont_path): obj_list) dir_list.append((path, do_stat(path).st_mtime)) + sleep() return ContainerDetails(bytes_used, object_count, obj_list, dir_list) |