Thread: pgpool 2.5b2 released
Pgpool 2.5b2 supports "master slave mode" which can cope with master/slave replication softwares such as Slony-I. In this mode pgpool sends non SELECT queries to master only. SELECTs are load balanced by pgpool. Other features of 2.5b2 include: - ability to add timestamp to each log entry - control to whether cache connection info or not pgpool 2.5b2 is available at: http://pgfoundry.org/projects/pgpool/ Enjoy, -- Tatsuo Ishii
Tatsuo Ishii wrote: > Pgpool 2.5b2 supports "master slave mode" which can cope with > master/slave replication softwares such as Slony-I. In this mode > pgpool sends non SELECT queries to master only. SELECTs are load > balanced by pgpool. > > Other features of 2.5b2 include: > - ability to add timestamp to each log entry > - control to whether cache connection info or not > > pgpool 2.5b2 is available at: > http://pgfoundry.org/projects/pgpool/ Wow, that is great! I know Jan was waiting for this. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073
From: "Tatsuo Ishii" <t-ishii@sra.co.jp> > Pgpool 2.5b2 supports "master slave mode" which can cope with > master/slave replication softwares such as Slony-I. In this mode > pgpool sends non SELECT queries to master only. SELECTs are load > balanced by pgpool. Sounds good! Does it attempt any interaction with Slony when it detects a failure of the master? It would seem a pity to have pgpool watching the pair to detect failure but having to have a separate watcher process to tell Slony to failover. Julian Scarfe
> > Pgpool 2.5b2 supports "master slave mode" which can cope with > > master/slave replication softwares such as Slony-I. In this mode > > pgpool sends non SELECT queries to master only. SELECTs are load > > balanced by pgpool. > > Sounds good! Thanks. Yesterday I have put offcial release of pgpool 2.5 on pgfoundry.org. The news release has not come up yet on the top page of pgfoundry.org for some reason I don't know, but I hope it will appear soon. pgpool 2.5 has the capabilty to perform periodical health checking to PostgreSQL. > Does it attempt any interaction with Slony when it detects a failure of the > master? It would seem a pity to have pgpool watching the pair to detect > failure but having to have a separate watcher process to tell Slony to > failover. If pgpool detects PostgreSQL failure, Slony should detect it as well, no? -- Tatsuo Ishii
> > Does it attempt any interaction with Slony when it detects a failure of the > > master? It would seem a pity to have pgpool watching the pair to detect > > failure but having to have a separate watcher process to tell Slony to > > failover. > > If pgpool detects PostgreSQL failure, Slony should detect it as well, no? Others are much better place to answer this, but my understanding is that Slony does not do so: http://gborg.postgresql.org/project/slony1/genpage.php?howto_overview "What Slony-I is not: Slony-I is not a network management system. Slony-I does not have any functionality within it to detect a node failure, or automatically promote a node to a master or other data origin. Slony-I is not multi-master; it's not a connection broker, and it doesn't make you coffee and toast in the morning." > pgpool 2.5 has the capabilty to perform periodical health checking to > PostgreSQL. Since pgpool has this capability, how about including a hook that allows a script to be run when pgpool detects a problem with the master? That would allow action to be taken to investigate further and, if required, switchover or failover and promote the slave to master. Julian Scarfe
On 2/2/2005 11:57 AM, Bruce Momjian wrote: > Tatsuo Ishii wrote: >> Pgpool 2.5b2 supports "master slave mode" which can cope with >> master/slave replication softwares such as Slony-I. In this mode >> pgpool sends non SELECT queries to master only. SELECTs are load >> balanced by pgpool. >> >> Other features of 2.5b2 include: >> - ability to add timestamp to each log entry >> - control to whether cache connection info or not >> >> pgpool 2.5b2 is available at: >> http://pgfoundry.org/projects/pgpool/ > > Wow, that is great! I know Jan was waiting for this. > This is incredible! I definitely did wait for this and some people at the Solutions Linux here in Paris did ask about it. Too bad that I didn't read this earlier, it would have made their day. Thank you Tatsuo. I will check this out when I'm back home. Jan -- #======================================================================# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #================================================== JanWieck@Yahoo.com #
>>pgpool 2.5 has the capabilty to perform periodical health checking to >>PostgreSQL. >> >> > >Since pgpool has this capability, how about including a hook that allows a >script to be run when pgpool detects a problem with the master? That would >allow action to be taken to investigate further and, if required, switchover >or failover and promote the slave to master. > > Because it is not the place of a connection management software to failover. Failover should happen only if: 1. You have a specific set of criteria that is matched via a network/system management software. 2. You do it yourself. The last thing in the world you need is to fail over to a slave because somebody accidently tripped over a network cord. Sincerely, Joshua D. Drake >Julian Scarfe > > > >---------------------------(end of broadcast)--------------------------- >TIP 6: Have you searched our list archives? > > http://archives.postgresql.org > > -- Command Prompt, Inc., home of Mammoth PostgreSQL - S/ODBC and S/JDBC Postgresql support, programming shared hosting and dedicated hosting. +1-503-667-4564 - jd@commandprompt.com - http://www.commandprompt.com PostgreSQL Replicator -- production quality replication for PostgreSQL
Attachment
From: "Joshua D. Drake" <jd@commandprompt.com> > >Since pgpool has this capability, how about including a hook that allows a > >script to be run when pgpool detects a problem with the master? That would > >allow action to be taken to investigate further and, if required, switchover > >or failover and promote the slave to master. > > > > > Because it is not the place of a connection management software > to failover. Failover should happen only if: > > 1. You have a specific set of criteria that is matched via a network/system > management software. > > 2. You do it yourself. I'm not suggesting that it's the place of pgpool to *force* a failover. I am suggesting that one of the criteria that is likely to be useful is the inability to connect to the master, and that's something that pgpool, apparently, detects. It seems unnecessary to use completely different failure-detection mechanisms for the purpose of failover to those used for the connection management. So all I'm looking for is a way for pgpool to shout if it detects a failure. That could initiate the investigation of the other criteria required for failover. > The last thing in the world you need is to fail over to a slave because > somebody accidently tripped over a network cord. In our application, that's *exactly* what we need. We have a database that receives data in a fairly continuous stream. If the datastream cannot be written to the database, the database becomes worse than useless quite rapidly. We need the ability to switchover or failover to another node as master as soon as possible, to allow the datastream to be written to the other node. We'll rebuild the "failed" master later, if necessary. But if the failover doesn't happen promptly, we might as well rebuild the whole cluster. Julian Scarfe
> I'm not suggesting that it's the place of pgpool to *force* a failover. I > am suggesting that one of the criteria that is likely to be useful is the > inability to connect to the master, and that's something that pgpool, > apparently, detects. It seems unnecessary to use completely different > failure-detection mechanisms for the purpose of failover to those used for > the connection management. > > So all I'm looking for is a way for pgpool to shout if it detects a failure. > That could initiate the investigation of the other criteria required for > failover. It's pretty easy. See main.c:failover_handler() for more details. -- Tatsuo Ishii > > The last thing in the world you need is to fail over to a slave because > > somebody accidently tripped over a network cord. > > In our application, that's *exactly* what we need. We have a database that > receives data in a fairly continuous stream. If the datastream cannot be > written to the database, the database becomes worse than useless quite > rapidly. We need the ability to switchover or failover to another node as > master as soon as possible, to allow the datastream to be written to the > other node. We'll rebuild the "failed" master later, if necessary. But if > the failover doesn't happen promptly, we might as well rebuild the whole > cluster. > > Julian Scarfe > > > > ---------------------------(end of broadcast)--------------------------- > TIP 5: Have you checked our extensive FAQ? > > http://www.postgresql.org/docs/faq >
After takin a swig o' Arrakan spice grog, julian@avbrief.com ("Julian Scarfe") belched out:
> So all I'm looking for is a way for pgpool to shout if it detects a
> failure. That could initiate the investigation of the other
> criteria required for failover.
_There_ lies the one change that is needed. Given that, some outside
'oracle' can be used to decide if it's appropriate to do a FAILOVER.
It's quite important for this not to be deeply embedded in pgpool...
--
let name="cbbrowne" and tld="ntlug.org" in name ^ "@" ^ tld;;
http://linuxdatabases.info/info/slony.html
PALINDROME spelled backwards is EMORDNILAP.
Hello, Command Prompt develops a lot of web based applications. One of the problems with the web is you can't really pre determine a lot of tasks because of the statelessness. It would be nice if you could tell pgpool to perform certain queries on initialization of a new connection. This would allow web based applications to take advantange of several nice features including: cursors for more than just the single session global variables preloaded indexes for performance in ram etc... Thoughts? Sincerely, Joshua D. Drake -- Command Prompt, Inc., home of Mammoth PostgreSQL - S/ODBC and S/JDBC Postgresql support, programming shared hosting and dedicated hosting. +1-503-667-4564 - jd@commandprompt.com - http://www.commandprompt.com PostgreSQL Replicator -- production quality replication for PostgreSQL
Attachment
> After takin a swig o' Arrakan spice grog, julian@avbrief.com ("Julian
Scarfe") belched out:
> > So all I'm looking for is a way for pgpool to shout if it detects a
> > failure. That could initiate the investigation of the other
> > criteria required for failover.
>
> _There_ lies the one change that is needed. Given that, some outside
> 'oracle' can be used to decide if it's appropriate to do a FAILOVER.
>
> It's quite important for this not to be deeply embedded in pgpool...
I think we have a consensus on that. So what's the most sensible mechanism
for the "shout". Since I posted my original question, I realized that
pgpool notes a failure of either master or slave in its log. Would we want
something more proactive?
Julian
>> Tatsuo Ishii wrote: >> >>> Pgpool 2.5b2 supports "master slave mode" which can cope with >>> master/slave replication softwares such as Slony-I. In this mode >>> pgpool sends non SELECT queries to master only. SELECTs are load >>> balanced by pgpool. First of all, thanks so much Tatsuo for such a great project! > This is incredible! I definitely did wait for this and some people at > the Solutions Linux here in Paris did ask about it. Too bad that I > didn't read this earlier, it would have made their day. Don't worry! On Feb 2nd, when we saw both Tatsuo did move his pgpool to pgfoundry, I downloaded the tarball, and played all the night with it, so on Feb 3rd (yep, last day, too bad), I gave back the info all the day on it to the people with the kind of problems pgpool and/or slony can resolve. Cheers, -- Jean-Paul Argudo www.PostgreSQLFr.org
# julian@avbrief.com / 2005-02-06 09:34:53 -0000:
> Since I posted my original question, I realized that pgpool notes a
> failure of either master or slave in its log. Would we want something
> more proactive?
snmp?
--
If you cc me or remove the list(s) completely I'll most likely ignore
your message. see http://www.eyrie.org./~eagle/faqs/questions.html
> Hello, > > Command Prompt develops a lot of web based applications. > One of the problems with the web is you can't really > pre determine a lot of tasks because of the statelessness. > > It would be nice if you could tell pgpool to perform certain > queries on initialization of a new connection. This would > allow web based applications to take advantange of several > nice features including: > > cursors for more than just the single session > global variables > preloaded indexes for performance in ram Yes, I thought about that too. Probably we need two kinds of initializations: 1) initilization for each new connection (as you requested) 2) initilization for each new client connection I think pgpool could do both of them since it has already done series of queries in closing client connections. -- Tatsuo Ishii
>>>>> "TI" == Tatsuo Ishii <t-ishii@sra.co.jp> writes: TI> Yes, I thought about that too. Probably we need two kinds of TI> initializations: TI> 1) initilization for each new connection (as you requested) TI> 2) initilization for each new client connection I think this latter one is important. Take the case where I do some "set FOO BAR" values to alter the query (eg, statement timeout or sequence scan disable) which should be reset on new client connection when it re-uses a backend connection. Or is this magically handled already? -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Vivek Khera, Ph.D. Khera Communications, Inc. Internet: khera@kciLink.com Rockville, MD +1-301-869-4449 x806 AIM: vivekkhera Y!: vivek_khera http://www.khera.org/~vivek/
> >>>>> "TI" == Tatsuo Ishii <t-ishii@sra.co.jp> writes: > > TI> Yes, I thought about that too. Probably we need two kinds of > TI> initializations: > > TI> 1) initilization for each new connection (as you requested) > > TI> 2) initilization for each new client connection > > I think this latter one is important. Take the case where I do some > "set FOO BAR" values to alter the query (eg, statement timeout or > sequence scan disable) which should be reset on new client connection > when it re-uses a backend connection. > > Or is this magically handled already? Yes. pgpool issues "RESET ALL" at the end of each client connection. See "reset_query_list" directive. -- Tatsuo Ishii
Roman Neuhauser wrote: > # julian@avbrief.com / 2005-02-06 09:34:53 -0000: > >>Since I posted my original question, I realized that pgpool notes a >>failure of either master or slave in its log. Would we want something >>more proactive? > > > snmp? Send out a NOTIFY that can be LISTENed for. -- Richard Huxton Archonet Ltd