Support separate replication ip address:
- Added new function in utils. This function provides ability
to select separate IP address for replication service.
- Db_replicator and object replicators were changed.
Replication process uses new function now.
Replication network parameters:
- Replication network fields (replication_ip, replication_port)
support was added to device dictionary in swift-ring-builder script.
- Changes were made to support new fields in search, show and set_info
functions.
Implementation of replication servers:
- Separate replication servers use the same code as normal replication
servers, but with replication_server parameter = True. When using a
separate replication network, the non-replication servers set
replication_server = False. When there is no separate replication
network (the default case), replication_server is not included in the config.
DocImpact
Change-Id: Ie9af5bdcdf9241c355e36053ca4adfe49dc35bd0
Implements: blueprint dedicated-replication-network
roundrobin_datadirs was returning any .db file at any depth in the
accounts/containers structure. Since xfs corruption can cause such
files to appear in odd places at times (only happened on one drive of
ours so far, but still...), I've refactored this function to only
return .db files at the proper depth.
Change-Id: Id06ef6584941f8a572e286f69dfa3d96fe451355
When a db is reclaimed it removes the hash dir the db files are in,
but it does not try to remove the parent suffix dir though it might
be empty now. This eventually leads to a bunch of empty suffix dirs
lying around. This patch fixes that by attempting to remove the
parent suffix dir after a hash dir reclamation.
Here's a quick script to see how bad a given drive might be:
import os, os.path, sys
if len(sys.argv) != 2:
sys.exit('%s <mount-point>' % sys.argv[0])
in_use = 0
empty = 0
containers = os.path.join(sys.argv[1], 'containers')
for p in os.listdir(containers):
partition = os.path.join(containers, p)
for s in os.listdir(partition):
suffix = os.path.join(partition, s)
if os.listdir(suffix):
in_use += 1
else:
empty += 1
print in_use, 'in use,', empty, 'empty,', '%.02f%%' % (
100.0 * empty / (in_use + empty)), 'empty'
And here's a quick script to clean up a drive:
NOTE THAT I HAVEN'T ACTUALLY RUN THIS ON A LIVE NODE YET!
import errno, os, os.path, sys
if len(sys.argv) != 2:
sys.exit('%s <mount-point>' % sys.argv[0])
containers = os.path.join(sys.argv[1], 'containers')
for p in os.listdir(containers):
partition = os.path.join(containers, p)
for s in os.listdir(partition):
suffix = os.path.join(partition, s)
try:
os.rmdir(suffix)
except OSError, err:
if err.errno not in (errno.ENOENT, errno.ENOTEMPTY):
print err
Change-Id: I2e6463a4cd40597fc236ebe3e73b4b31347f2309
To tell when replication for a device has finished, it's important to
know when the replicator is removing objects. This was previously
handled for the object-replicator
(object-replicator.partition.delete.count.<device> and
object-replicator.partition.update.count.<device> metrics) but not the
account and container replicators.
This patch extends the existing DB removal count metrics to make them
per-device. The new metrics are:
account-replicator.removes.<device>
container-replicator.removes.<device>
There's also a bonus refactoring and increased test coverage of the DB
replicator code.
Change-Id: I2067317d4a5f8ad2a496834147954bdcdfc541c1
