Thanks Akshesh. Disconnecting participants is one way to achieve the goal.
But it could impact the all replicas on that participant.
I would suggest Imran to disable the partition for that instance. So it
will trigger it from Master -> Slave. Then you can throw exception during
the process to
Hi Imran
I am not sure if you can mark it as ERROR or not (since the transition S->M
has already completed successfully) but if you just care about this replica
going down and spawning a master on another node you can just stop this
replica by stopping this process or *disconnecting
Hi Lei,
The partition is already in M state as transition was completed quite
sometime back. Now, only 1 partition in the participant needs to be moved
to ERROR state as some I/O exception has happened. So is there any API that
the participant could call to mark this particular partition in ERROR
Hi, Imran
You can simply throw out an exception from your M->S state transition,
then Helix will mark it as ERROR.
Lei
On Fri, Jun 12, 2020 at 4:41 AM Imran S. Shah wrote:
> Hello
>
> I have implemented a standard Master-Slave model. Now let's say my
> partitions are in Master state, but
Shah
Gesendet: Freitag, 12. Juni 2020 13:42
An: user@helix.apache.org
Betreff: Question: How to transition from Master to Error
Hello
I have implemented a standard Master-Slave model. Now let's say my partitions
are in Master state, but however due to some errors (out of disk space for
