Thread: Slony in pgAdmin

As I might need Slony in my project, I decided to look
at it more closely.

What's with Slony-I support in pgAdmin?
Is someone working on it?

When I use right-click New Object->Slony-I Cluster,
window appears, but on the bottom there is string:
"Slony-I creation scripts not available; only joining possible."

I tried to look into source code, but I'm not sure where to start,
and what should be done.
Is there any plan of implementing Slony support in pgAdmin?
For example:
1. Create code for creating, deleting and managing clusters.
2. -''- -''- nodes.
3. -''- -''- sets, adding tables and sequences.
4. etc.

And, if I've seen correctly, pgAdmin doesn't use slonik,
and creates it's own queries to do things in cluster.
Is it right?

I'd like to start playing a little with code.
If I start, I hope I won't get into someone's area,
so if someone is working at Slony support,
please let me know, and tell what's left to play with.

I think that the best place to start would be
cluster code, located in dlgRepCluster and slCluster.
Is that right?

And one last question.
There is no Slony documentation in pgAdmin, isn't?

--
Tomasz Rybak <bogomips@post.pl>

> -----Original Message-----
> From: pgadmin-hackers-owner@postgresql.org
> [mailto:pgadmin-hackers-owner@postgresql.org] On Behalf Of
> Tomasz Rybak
> Sent: 18 September 2005 21:58
> To: pgAdmin Hackers
> Subject: [pgadmin-hackers] Slony in pgAdmin
>
> As I might need Slony in my project, I decided to look
> at it more closely.
>
> What's with Slony-I support in pgAdmin?
> Is someone working on it?

Absolutely. We (well, Andreas anyway) has to give a presentation of it
to a large usergroup at the end of October, so it's definitely being
worked on. We've spent a fair bit on plane tickets etc. so this one's on
a real deadline!

> When I use right-click New Object->Slony-I Cluster,
> window appears, but on the bottom there is string:
> "Slony-I creation scripts not available; only joining possible."

The Slony SQL scripts need to be somewhere that pgAdmin can find them. I
forget where off-hand, but watch the list because I was going to email
Andreas about it this morning anyway!

> I tried to look into source code, but I'm not sure where to start,
> and what should be done.
> Is there any plan of implementing Slony support in pgAdmin?
> For example:
> 1. Create code for creating, deleting and managing clusters.
> 2. -''- -''- nodes.
> 3. -''- -''- sets, adding tables and sequences.
> 4. etc.

All that was certainly there for Slony 1.0 last time I tried it. Andreas
has been working on support for 1.1/1.2 most recently.

> And, if I've seen correctly, pgAdmin doesn't use slonik,
> and creates it's own queries to do things in cluster.
> Is it right?

Yes.

> I'd like to start playing a little with code.
> If I start, I hope I won't get into someone's area,
> so if someone is working at Slony support,
> please let me know, and tell what's left to play with.

To be honest, I would keep out of there for the moment unless Andreas
has anything specific he'd like some help with.

> I think that the best place to start would be
> cluster code, located in dlgRepCluster and slCluster.
> Is that right?

Yes.

> And one last question.
> There is no Slony documentation in pgAdmin, isn't?

No. That should be failry easy to add assuming the Slony docs have the
make htmlhelp target that pg does. I doubt Andreas has done anything on
that if you want to look at it.

Regards, Dave.

Dave Page wrote:

>
>
>>What's with Slony-I support in pgAdmin?
>>Is someone working on it?
>>
>>
>
>Absolutely. We (well, Andreas anyway) has to give a presentation of it
>to a large usergroup at the end of October, so it's definitely being
>worked on. We've spent a fair bit on plane tickets etc. so this one's on
>a real deadline!
>
>
I want to present it working, and have it released working in V1.4
(concurrent with pgsql 8.1), so that's the time frame.

>>When I use right-click New Object->Slony-I Cluster,
>>window appears, but on the bottom there is string:
>>"Slony-I creation scripts not available; only joining possible."
>>
>>
See Options. It used to be located over the path, has its setting now.

>>For example:
>>1. Create code for creating, deleting and managing clusters.
>>2. -''- -''- nodes.
>>3. -''- -''- sets, adding tables and sequences.
>>4. etc.
>>
>>
1), 2), 3) is working, as well as creating pgsql objects
replication-aware, i.e. using Slony-I.
4) is e.g. failover, which has to be coded yet, and some more. These
should cover the error-prone cases, and should receive a fair amount
review because taken from the slony mailing list quite a lot may go
wrong with that; lot of testing required.

>
>
>
>>I'd like to start playing a little with code.
>>If I start, I hope I won't get into someone's area,
>>so if someone is working at Slony support,
>>please let me know, and tell what's left to play with.
>>
>>
>
>
>
If you want to help, you can test, test, test.

>
>
>>And one last question.
>>There is no Slony documentation in pgAdmin, isn't?
>>
>>
>
>No. That should be failry easy to add assuming the Slony docs have the
>make htmlhelp target that pg does. I doubt Andreas has done anything on
>that if you want to look at it.
>
>
Unfortunately we'd need to write our own doc, since things are handled a
bit more abstract than with slonik. But we certainly should include the
whole slony doc as subtree in the main pgAdmin doc.
I don't know if anybody has ever tested to generate htmlhelp from slony,
whould be a good start to see if it works and can be compiled to a .CHM
with M$'s htmlhelp compiler, which would be a good start for further
integration.

Regards,
Andreas

Dnia 19-09-2005, pon o godzinie 16:32 +0200, Andreas Pflug napisał(a):
> Dave Page wrote:
> >>What's with Slony-I support in pgAdmin?
> >>Is someone working on it?
> >Absolutely. We (well, Andreas anyway) has to give a presentation of it
> >to a large usergroup at the end of October, so it's definitely being
> >worked on. We've spent a fair bit on plane tickets etc. so this one's on
> >a real deadline!
> I want to present it working, and have it released working in V1.4
> (concurrent with pgsql 8.1), so that's the time frame.
>

OK.

> >>When I use right-click New Object->Slony-I Cluster,
> >>window appears, but on the bottom there is string:
> >>"Slony-I creation scripts not available; only joining possible."
> >>
> >>
> See Options. It used to be located over the path, has its setting now.

Thanks, I found it and now have more options.
However I've installed Slony-I 1.1.0, so few thinks may don't work.

>
> >>For example:
> >>1. Create code for creating, deleting and managing clusters.
> >>2. -''- -''- nodes.
> >>3. -''- -''- sets, adding tables and sequences.
> >>4. etc.
> >>
> >>
> 1), 2), 3) is working, as well as creating pgsql objects
> replication-aware, i.e. using Slony-I.
> 4) is e.g. failover, which has to be coded yet, and some more. These
> should cover the error-prone cases, and should receive a fair amount
> review because taken from the slony mailing list quite a lot may go
> wrong with that; lot of testing required.
>
> >
> >
> >
> >>I'd like to start playing a little with code.
> >>If I start, I hope I won't get into someone's area,
> >>so if someone is working at Slony support,
> >>please let me know, and tell what's left to play with.
> >>
> >>
> >
> >
> >
> If you want to help, you can test, test, test.
>

OK. I'll try.

First note.
I created Slony cluster using slonik, and when I open database,
options from Tools->Replication aren't active (are grayed out).

> >
> >
> >>And one last question.
> >>There is no Slony documentation in pgAdmin, isn't?
> >>
> >>
> >
> >No. That should be failry easy to add assuming the Slony docs have the
> >make htmlhelp target that pg does. I doubt Andreas has done anything on
> >that if you want to look at it.
> >
> >
> Unfortunately we'd need to write our own doc, since things are handled a
> bit more abstract than with slonik. But we certainly should include the
> whole slony doc as subtree in the main pgAdmin doc.
> I don't know if anybody has ever tested to generate htmlhelp from slony,
> whould be a good start to see if it works and can be compiled to a .CHM
> with M$'s htmlhelp compiler, which would be a good start for further
> integration.
>

I don't have Windows, so here I cannot help.

--
Tomasz Rybak <bogomips@post.pl>

Dnia 20-09-2005, wto o godzinie 19:45 +0200, Tomasz Rybak napisał(a):
<cut>
> Thanks, I found it and now have more options.
> However I've installed Slony-I 1.1.0, so few thinks may don't work.
>
<cut>
> First note.
> I created Slony cluster using slonik, and when I open database,
> options from Tools->Replication aren't active (are grayed out).

I noticed that new subtree appeared, and when I click on
Replication->ClasterName options are active.

However, Failover option doesn't work - when I click on it,
nothing happens, and in slon's log also nothing appears.
Restart Node works and Upgrade Node works (at least it writes
"Selected cluster doesn't contain newer software.".

--
Tomasz Rybak <bogomips@post.pl>

Tomasz Rybak wrote:

>
> However, Failover option doesn't work - when I click on it,
> nothing happens, and in slon's log also nothing appears.

Not surprising, as I wrote it's not implemented yet; unsure about the
other functions.

> Restart Node works and Upgrade Node works (at least it writes
> "Selected cluster doesn't contain newer software.".

I'm not sure if update is alread complete; AFAIR no. I'll keep you cc-ed
  about coding progress.

Regards,
Andreas

I am very interested in using pgadmin to configure my Slony-I.
I have the same kind of problem.
But I would like to know if it is possible to use pgadmin running under Windows to configure a slony on a Suse machine.

---

Sent from the PostgreSQL - pgadmin hackers forum at Nabble.com:
Re: Slony in pgAdmin