Move cluster utilities out of contrib

One comment:

pgxc_ctl and pgxc_clean were in contrib because they don't provide all the possible means of cluster configuration/operation and cleanup.

For the usage which they don't support yet, we need to keep expanding them. For example, pgxc_ctl does not support more than one slave, as well as cascaded slave, It does not support fallback using older database image either. People may want to have more handy backup and restoration means through pgxc_ctl.

I hope move to /src/bin means to keep improving them including the above issues.

One comment:

pgxc_ctl and pgxc_clean were in contrib because they don't provide all
the possible means of cluster configuration/operation and cleanup.

For the usage which they don't support yet, we need to keep expanding
them. For example, pgxc_ctl does not support more than one slave,
as well as cascaded slave, It does not support fallback using older
database image either. People may want to have more handy backup
and restoration means through pgxc_ctl.

I hope move to /src/bin means to keep improving them including the above issues.

Regards;

Koichi Suzuki

2014-10-08 4:36 GMT+09:00 mason_s mason_s@users.sf.net:

---

[tickets:#33] Move cluster utilities out of contrib

Status: open
Milestone: 9.2rc
Created: Tue Oct 07, 2014 07:36 PM UTC by mason_s
Last Updated: Tue Oct 07, 2014 07:36 PM UTC
Owner: mason_s

The contrib source directory contains optional modules and includes
pgxc_ctl, pgxc_clean and pgxc_monitor. The first two in particular are very
important for management, should always be built and should not be
considered optional. This will also help with questions from people asking
about where those utilities are.

We should move these instead to src/bin, to always be compiled when building
from source.

---

Sent from sourceforge.net because you indicated interest in
https://sourceforge.net/p/postgres-xl/tickets/33/

To unsubscribe from further messages, please visit
https://sourceforge.net/auth/subscriptions/

There had been many people struggling with manual configurations of the cluster, and pgxc_ctl helped greatly in getting people up and running sooner and starting, stopping, etc. (and saved us some time answering questions on the mailing list).

Pavan has addressed some bugs in recent weeks. I agree that we need to continue improving it, particularly additional data node slaves as you say, as well as the ability to failback.