Commit dd556379 authored by Julien Muchembled's avatar Julien Muchembled

Comment about backup limitations

parent 8832bd42
......@@ -85,6 +85,11 @@ RC - Review output of pylint (CODE)
Currently, storage presence is broadcasted to client nodes too early, as
the storage node would refuse them until it has only up-to-date data (not
only up-to-date cells, but also a partition table and node states).
- In backup mode, 2 simultaneous replication should be possible so that:
- outdated cells does not block backup for too long time
- constantly modified partitions does not prevent outdated cells to
replicate
Current behaviour is undefined and the above 2 scenarios may happen.
- Create a specialized PartitionTable that know the database and replicator
to remove duplicates and remove logic from handlers (CODE)
- Consider insert multiple objects at time in the database, with taking care
......
......@@ -141,6 +141,9 @@ class BackupApplication(object):
tid = app.backup_tid
# Wait for non-primary partitions to catch up,
# so that all UP_TO_DATE cells are really UP_TO_DATE.
# XXX: Another possibility could be to outdate such cells, and
# they would be quickly updated at the beginning of the
# RUNNING phase. This may simplify code.
# Any unfinished replication from upstream will be truncated.
while pt.getCheckTid(xrange(pt.getPartitions())) < tid:
poll(1)
......
......@@ -227,6 +227,8 @@ class Replicator(object):
return
app = self.app
# Choose a partition with no unfinished transaction if possible.
# XXX: When leaving backup mode, we should only consider UP_TO_DATE
# cells when leaving backup mode.
for offset in self.replicate_dict:
if not self.partition_dict[offset].max_ttid:
break
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment