I am having a project that will be using postgresql for HA (high
availablity) and load balancing.

Being search for a while, here is what I found,


basically for HA, it syncs "xlog" between master and slave, you can use
native tool in crontab, or use 3rd party tools, like pluster, pgpool,
plproxy,slony-i

Am I right?


So for which platform to use, I like to hear some advice.

Linux, windows or Solaris (SPARC)

Thanks for your time.

steeles <steeles (AT) gmail (DOT) com> wrote:
Spend some time to define your service level, i.e. what kind of failures
can you tolerate, how long is the acceptable downtime for each,
how much data loss is acceptable in each of these situations etc.

You'll need that to decide.

Then decide upon a technique that you want to use to fulfill that.

I think that that question has little to do with high availibility.

Choose the one that you or your administrators are most familiar with.

Of course excluding Windows. You wouldn't use Windows anywhere else
that on an end user desktop.

Yours,
Laurenz Albe

steeles <steeles (AT) gmail (DOT) com> wrote:
Spend some time to define your service level, i.e. what kind of failures
can you tolerate, how long is the acceptable downtime for each,
how much data loss is acceptable in each of these situations etc.

You'll need that to decide.

Then decide upon a technique that you want to use to fulfill that.

I think that that question has little to do with high availibility.

Choose the one that you or your administrators are most familiar with.

Of course excluding Windows. You wouldn't use Windows anywhere else
that on an end user desktop.

Yours,
Laurenz Albe

steeles <steeles (AT) gmail (DOT) com> wrote:
Spend some time to define your service level, i.e. what kind of failures
can you tolerate, how long is the acceptable downtime for each,
how much data loss is acceptable in each of these situations etc.

You'll need that to decide.

Then decide upon a technique that you want to use to fulfill that.

I think that that question has little to do with high availibility.

Choose the one that you or your administrators are most familiar with.

Of course excluding Windows. You wouldn't use Windows anywhere else
that on an end user desktop.

Yours,
Laurenz Albe

steeles <steeles (AT) gmail (DOT) com> wrote:
Spend some time to define your service level, i.e. what kind of failures
can you tolerate, how long is the acceptable downtime for each,
how much data loss is acceptable in each of these situations etc.

You'll need that to decide.

Then decide upon a technique that you want to use to fulfill that.

I think that that question has little to do with high availibility.

Choose the one that you or your administrators are most familiar with.

Of course excluding Windows. You wouldn't use Windows anywhere else
that on an end user desktop.

Yours,
Laurenz Albe

steeles <steeles (AT) gmail (DOT) com> wrote:
Spend some time to define your service level, i.e. what kind of failures
can you tolerate, how long is the acceptable downtime for each,
how much data loss is acceptable in each of these situations etc.

You'll need that to decide.

Then decide upon a technique that you want to use to fulfill that.

I think that that question has little to do with high availibility.

Choose the one that you or your administrators are most familiar with.

Of course excluding Windows. You wouldn't use Windows anywhere else
that on an end user desktop.

Yours,
Laurenz Albe

steeles <steeles (AT) gmail (DOT) com> wrote:
Spend some time to define your service level, i.e. what kind of failures
can you tolerate, how long is the acceptable downtime for each,
how much data loss is acceptable in each of these situations etc.

You'll need that to decide.

Then decide upon a technique that you want to use to fulfill that.

I think that that question has little to do with high availibility.

Choose the one that you or your administrators are most familiar with.

Of course excluding Windows. You wouldn't use Windows anywhere else
that on an end user desktop.

Yours,
Laurenz Albe

steeles <steeles (AT) gmail (DOT) com> wrote:
Spend some time to define your service level, i.e. what kind of failures
can you tolerate, how long is the acceptable downtime for each,
how much data loss is acceptable in each of these situations etc.

You'll need that to decide.

Then decide upon a technique that you want to use to fulfill that.

I think that that question has little to do with high availibility.

Choose the one that you or your administrators are most familiar with.

Of course excluding Windows. You wouldn't use Windows anywhere else
that on an end user desktop.

Yours,
Laurenz Albe

steeles <steeles (AT) gmail (DOT) com> wrote:
Spend some time to define your service level, i.e. what kind of failures
can you tolerate, how long is the acceptable downtime for each,
how much data loss is acceptable in each of these situations etc.

You'll need that to decide.

Then decide upon a technique that you want to use to fulfill that.

I think that that question has little to do with high availibility.

Choose the one that you or your administrators are most familiar with.

Of course excluding Windows. You wouldn't use Windows anywhere else
that on an end user desktop.

Yours,
Laurenz Albe