#48999 create keep alive entry after on line init
Opened 3 years ago by lkrispen. Modified 3 months ago

In ticket 48445 the creation of a keep alive entry aftzer online init was removed because the creation i replica_enable_replication() led to inconsistencies between database and changelog.
But the creation of a keep alive entry directly after init is useful to have a csn for teh local replica in the RUV.
A safer place to add it is in the attached patch


Metadata Update from @lkrispen:
- Issue set to the milestone: 1.3.5.14

2 years ago

Metadata Update from @lkrispen:
- Issue assigned to lkrispen

2 years ago

To echo the voice of Mark "You have an indentation problem in this patch".

Otherwise, the patch seems to make sense. I assume that this keep alive addition is atomic relative to the operation of the init? IE we won't see:

init
incremental
keep alive

In that order? the init and keep alive are not able to be seperated?

Metadata Update from @firstyear:
- Issue close_status updated to: None

2 years ago

Metadata Update from @mreynolds:
- Issue set to the milestone: 1.3.7 backlog (was: 1.3.5.14)

2 years ago

Metadata Update from @mreynolds:
- Issue set to the milestone: 1.3.7.0 (was: 1.3.7 backlog)

2 years ago

Metadata Update from @mreynolds:
- Custom field reviewstatus adjusted to None
- Issue set to the milestone: 1.4.2 (was: 1.3.7.0)

3 months ago

I think it is, but the patch has to be rebased, correctly indented and tested again.

and I would like t get the opinion of @tbordaz

The patch allows a consumer (configured as a master) to create (if it did not exist before) the consumer's keep alive entry. It creates it at the end of a replication session at a time we are sure the changelog exists and will not be reset. It allows a consumer to have RUVelement with csn in the RUV at the first incoming replication session.

Without that patch, the consumer waits to behave as a master and send data to create the RUVelement. So the RUV on the consumer may have empty RUVelement until it receives a direct update.

The patch looks good to me to obtain a non empty RUVelement. ACK

Login to comment on this ticket.

Metadata