On Thu, Jan 16, 2014 at 03:29:49PM +0100, Michelle Sullivan wrote: > Yup confirmed... the delete process is being sent to the masters > sequentially... this is likely to be a cause of some of the performance > issues. Surely as bucardo holds a connection to each DB independantly > it should run the deletes on the same table in parallel to all hosts in > the sync and then wait for them all to finish? before completing the > transaction - it would probably result in better performance all round..
The deletes should be happening asynchronously - in other words, all at once. See around line 8520 in git head. For each target, we send a list of rows to be deleted, with pg_async true (hopefully). Once they are all sent off, we wait for them all to finish, then start copying in new rows. If you can confirm that things are not happening asynchronously, it is probably a bug. Lemme commit a quick debug aid to help ... okay, it's in 4c88741a1d579f8c5c8b752eb9eb3a3f698a75d8 -- Greg Sabino Mullane [email protected] End Point Corporation PGP Key: 0x14964AC8
signature.asc
Description: Digital signature
_______________________________________________ Bucardo-general mailing list [email protected] https://mail.endcrypt.com/mailman/listinfo/bucardo-general
