Re: [HACKERS] 7.4 hot backup capabilities?
Austin Gonyou wrote: What facilities are/will be available for hot(online) backups with the 7.4 release? PITR, something else? TIA. pg_dump? Did you mean hot failover or hot backup? Postgresql does hot backup for a long time. Bye Shridhar ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [HACKERS] CVS with tag REL7_4_STABLE don't compile
> I have a problem with compile PostgreSQL from cvs with tag > REL7_4_STABLE > Error is 'ECPG_ARRAY_NONE is not declared' (in execute.c) My fault. extern.h wasn't committed and I didn't notice. Sorry. It should be fixed now. Michael -- Michael Meskes Email: Michael at Fam-Meskes dot De ICQ: 179140304, AIM/Yahoo: michaelmeskes, Jabber: [EMAIL PROTECTED] Go SF 49ers! Go Rhein Fire! Use Debian GNU/Linux! Use PostgreSQL! ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
Re: [HACKERS] 7.4 hot backup capabilities?
On Wed, 2003-11-12 at 01:00, Shridhar Daithankar wrote: > Austin Gonyou wrote: > > > What facilities are/will be available for hot(online) backups with the > > 7.4 release? PITR, something else? TIA. > > pg_dump? > > Did you mean hot failover or hot backup? Postgresql does hot backup for a long time. I was referring primarly to hot backup. We're migrating from Oracle to pgsql, and we wanted to know more about this. From our current understanding, pgsql's hot backup is more of an export, than say, a hot backup with redo-logs type mode.(i.e. oracle rman). Is there a similar utility? I didn't think that pg_dump did that functionality as an actual "backup". i.e. keeps track of block changes while IO is going on and ensures stuff in process gets backed up as well. > Bye >Shridhar -- Austin Gonyou <[EMAIL PROTECTED]> Coremetrics, Inc. ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [HACKERS] 7.4 hot backup capabilities?
Austin Gonyou wrote: On Wed, 2003-11-12 at 01:00, Shridhar Daithankar wrote: Austin Gonyou wrote: What facilities are/will be available for hot(online) backups with the 7.4 release? PITR, something else? TIA. pg_dump? Did you mean hot failover or hot backup? Postgresql does hot backup for a long time. I was referring primarly to hot backup. We're migrating from Oracle to pgsql, and we wanted to know more about this. From our current understanding, pgsql's hot backup is more of an export, than say, a hot backup with redo-logs type mode.(i.e. oracle rman). Is there a similar utility? I didn't think that pg_dump did that functionality as an actual "backup". i.e. keeps track of block changes while IO is going on and ensures stuff in process gets backed up as well. pg_dump takes a snapshot while it starts and it saves the snapshot consistently. Any ongoing transactions that are not yet committed are not visible to pg_dump anyways. Furthermore since the snapshot is consistent, pg_dump does not need redo logs etc. You can look into asynchronous replications if you want incremental backup. There is a gborg page for replication projects, http://gborg.postgresql.org/project/pgreplication/projdisplay.php pg_dump can be selective but I don't think it can be incremental. I hope I got your point correctly. HTH Shridhar ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [HACKERS] 7.4 hot backup capabilities?
Shridhar Daithankar <[EMAIL PROTECTED]> writes: > pg_dump can be selective but I don't think it can be incremental. > > I hope I got your point correctly. I think he's really talking about PITR, which will not be in 7.4. -Doug ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
Re: [HACKERS] NetBSD/acorn32
On Sun, Nov 09, 2003 at 09:51:31PM +0100, Peter Eisentraut wrote: > Patrick Welche writes: > > > PostgreSQL 7.5devel on arm-unknown-netbsdelf1.6ZE, compiled by GCC gcc (GCC) > > 3.3.2-nb1 > > Can you test 7.4? Several days of compiling later.. All 93 tests passed. with 7.4rc2 on NetBSD-1.6ZE/acorn32. Patrick ---(end of broadcast)--- TIP 8: explain analyze is your friend
[HACKERS] rpm support for 7.4 and beyond
I haven't seen any discussion on the topic, so thought I might start
one. We currently provide rpms for Red Hat 7.x, 8.x, and 9, and I'm
assuming that we'll have rpms for all of those for 7.4 once release time
rolls around. However, given that Red Hat is dropping support for 7.x
and 8.x version on Jan 1st, and Red Hat 9 next April, do the current rpm
builders forsee any issues with providing rpms for those platforms in
the future?
Also (and maybe someone from Red Hat can weigh in here) are there any
plans from Red Hat to release RHEL rpms for postgresql in the future,
and/or plans to make sure the community rpm builders would have access
to those platforms in order to build rpms against them?
Robert Treat
--
Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL
---(end of broadcast)---
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [EMAIL PROTECTED] so that your
message can get through to the mailing list cleanly
Re: [HACKERS] 7.4 hot backup capabilities?
Austin Gonyou wrote: On Wed, 2003-11-12 at 01:00, Shridhar Daithankar wrote: Austin Gonyou wrote: > What facilities are/will be available for hot(online) backups with the > 7.4 release? PITR, something else? TIA. pg_dump? Did you mean hot failover or hot backup? Postgresql does hot backup for a long time. I was referring primarly to hot backup. We're migrating from Oracle to pgsql, and we wanted to know more about this. From our current understanding, pgsql's hot backup is more of an export, than say, a hot backup with redo-logs type mode.(i.e. oracle rman). Is there a similar utility? I didn't think that pg_dump did that functionality as an actual "backup". i.e. keeps track of block changes while IO is going on and ensures stuff in process gets backed up as well. You might want to read section 1.5 of my Slony-I proposal http://developer.postgresql.org/~wieck/slony1.html It might take 2-3 months until we have the first BETA. I hope that will fit your migration projects time table. Jan -- #==# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #== [EMAIL PROTECTED] # ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [HACKERS] rpm support for 7.4 and beyond
Robert Treat <[EMAIL PROTECTED]> writes: > Also (and maybe someone from Red Hat can weigh in here) are there any > plans from Red Hat to release RHEL rpms for postgresql in the future, I can tell you that Red Hat is getting beat up regularly for having omitted Postgres (and MySQL!) from RHEL 3. If you are a paying RHEL customer, make sure you let them know you're not happy about it. Upper management keep changing their mind about how exactly they want to support these databases on RHEL --- well, okay, that's a business decision and outside my sphere as an engineer. But in the meantime the effective support is "none at all", and people have got to hold their feet to the fire about it. regards, tom lane ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
Re: [HACKERS] 7.4RC2 regression failur and not running stats collector process on Solaris
> > LOG: could not bind socket for statistics collector: Cannot assign requested > > address > > Hmm ... that's sure the problem, but what can we do about it? ISTM that > any non-broken system ought to be able to resolve "localhost". Actually > it's worse than that: your system resolved "localhost" and then refused Are we using an api that only returns nslookup responses and not /etc/hosts entries ? At least on AIX it looks like it. Andreas ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [HACKERS] 7.4RC2 regression failur and not running stats collector process on Solaris
"Zeugswetter Andreas SB SD" <[EMAIL PROTECTED]> writes: > Are we using an api that only returns nslookup responses and not > /etc/hosts entries ? At least on AIX it looks like it. We use getaddrinfo(), or if that doesn't exist gethostbyname(). If there's a problem of that ilk then it's those library routines' fault. But AFAICT Kiyoshi's problem is not that ... unless maybe localhost is incorrectly listed as something other than 127.0.0.1 in one of those sources? regards, tom lane ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [HACKERS] rpm support for 7.4 and beyond
Tom Lane wrote: Robert Treat <[EMAIL PROTECTED]> writes: Also (and maybe someone from Red Hat can weigh in here) are there any plans from Red Hat to release RHEL rpms for postgresql in the future, I can tell you that Red Hat is getting beat up regularly for having omitted Postgres (and MySQL!) from RHEL 3. If you are a paying RHEL customer, make sure you let them know you're not happy about it. Upper management keep changing their mind about how exactly they want to support these databases on RHEL --- well, okay, that's a business decision and outside my sphere as an engineer. But in the meantime the effective support is "none at all", and people have got to hold their feet to the fire about it. I just spent 15 minutes searching on the RH web site trying to locate a complete list of packages in the various RHEL personalities, with conspicuous lack of success. How anyone can make decisions about it without knowing exactly what is in it is beyond me. To answer Robert's original question, I suspect that there will be many corporate and individual users of RH 7.x, 8 and 9 for quite some time, and providing RPMs for these will be highly desirable. I currently have access to one of each of these (fairly vanilla installations, too), and can help out if need be while that lasts - I have no plans to upgrade any of them right now, and would only want to do so if some critical issue came up that forced it - we are behind a firewall so even remote root exploits might not make me upgrade. cheers andrew ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [HACKERS] rpm support for 7.4 and beyond
On Wed, 2003-11-12 at 10:23, Tom Lane wrote:
> Robert Treat <[EMAIL PROTECTED]> writes:
> > Also (and maybe someone from Red Hat can weigh in here) are there any
> > plans from Red Hat to release RHEL rpms for postgresql in the future,
>
> I can tell you that Red Hat is getting beat up regularly for having
> omitted Postgres (and MySQL!) from RHEL 3. If you are a paying RHEL
> customer, make sure you let them know you're not happy about it.
>
> Upper management keep changing their mind about how exactly they want
> to support these databases on RHEL --- well, okay, that's a business
> decision and outside my sphere as an engineer. But in the meantime the
> effective support is "none at all", and people have got to hold their
> feet to the fire about it.
>
well, i'm not now, but if they want me to become a paying RHEL customer
they better make me feel secure that there is going to be some support
for postgresql on those platforms, even if it's just a statement that
they'll give "Lamar & Company" free access to the flavors of RHEL to do
community rpm building.
Personally I don't find any of their current RHEL offerings that
enticing given that the extent of my support needs over the last year or
so involved downloading maybe two kernel patches and an ssh patch, but
I've got a number of Red Hat 7.3 servers I'll eventually have to
transition to *something* sooner or later...
Robert Treat
--
Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL
---(end of broadcast)---
TIP 8: explain analyze is your friend
Re: [HACKERS] rpm support for 7.4 and beyond
I wrote: I just spent 15 minutes searching on the RH web site trying to locate a complete list of packages in the various RHEL personalities, with conspicuous lack of success. How anyone can make decisions about it without knowing exactly what is in it is beyond me. However, you can see a list of its SRPMs here: http://distro.ibiblio.org/pub/Linux/distributions/redhat/enterprise/3/en/os/i386/SRPMS/ and presumably on other mirrors. it includes rh-postgresql-7.3.4-8.src.rpm (14.9Mb) and mysql-3.23.58-1.src.rpm (11.5 Mb) cheers andrew ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
[HACKERS] Successful proting report for 7.4beta4 on Irix and Altix
Dear All, I have compiled and tested Postgresql-7.4beta4 on SGI Irix 6.5.20 using the MIPS Pro 7.4 compilers and on SGI Altix using the gcc 2.96 compilers and both ports passed all tests within the limits of numerical accuracy. What is the URL for reporting successful ports? --Bob +-++ | Robert E. Bruccoleri, Ph.D. | email: [EMAIL PROTECTED]| | President, Congenomics Inc. | URL: http://www.congen.com/~bruc | | P.O. Box 314| Phone: 609 818 7251| | Pennington, NJ 08534|| +-++ ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
Re: [HACKERS] 7.4RC2 regression failur and not running stats collector process on Solaris
On Wed, Nov 12, 2003 at 10:32:38AM -0500, Tom Lane wrote: > "Zeugswetter Andreas SB SD" <[EMAIL PROTECTED]> writes: > > Are we using an api that only returns nslookup responses and not > > /etc/hosts entries ? At least on AIX it looks like it. > > We use getaddrinfo(), or if that doesn't exist gethostbyname(). > If there's a problem of that ilk then it's those library routines' > fault. But AFAICT Kiyoshi's problem is not that ... unless maybe > localhost is incorrectly listed as something other than 127.0.0.1 > in one of those sources? It might depend on settings in /etc/host.conf or /etc/nsswitch.conf or something too? You can ussually tell the lib to use the files or not. It's always a good idea to put localhost into dns too. Kurt ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [HACKERS] 7.4RC2 regression failur and not running stats collector process on Solaris
Kurt Roeckx <[EMAIL PROTECTED]> writes: > It's always a good idea to put localhost into dns too. Yeah, but "localhost" *is* resolving as something on Kiyoshi's machine, else a different error message would have appeared. I'm wondering just what it resolved to though --- maybe we should have made the error messages more verbose, or added a debug-level message to show what addresses are being tried. regards, tom lane ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [HACKERS] Proposal for a cascaded master-slave replication system
Hans-Jürgen Schönig wrote: Jan, First of all we really appreciate that this is going to be an Open Source project. There is something I wanted to add from a marketing point of view: I have done many public talks in the 2 years or so. There is one question people keep asking me: "How about the pgreplication project?". In every training course, at any conference people keep asking for synchronous replication. We have offered this people some async solutions which are already out there but nobody seems to be interested in having it (my person impression). People keep asking for a sync approach via email but nobody seems to care about an async approach. This does not mean that async is bad but we can see a strong demand for synchronous replication. Meanwhile we seem to be in a situation where PostgreSQL is rather competing against Oracle than against MySQL. In our case there are more people asking for Oracle -> Pg migration than for MySQL -> Pg. MySQL does not seem to be the great enemy because most people know that it is an inferior product anyway. What I want to point out is that some people want an alternative Oracle's Real Application Cluster. They want load balancing and hot failover. Even data centers asking for replication did not want to have an async approach in the past. Hans-Jürgen, we are well aware of the high demand for multi-master replication addressing load balancing and clustering. We have that need ourself as well and I plan to work on a follow-up project as soon as Slony-I is released. But as of now, we see a higher priority for a reliable master slave system that includes the cascading and backup features described in my concept. There are a couple of different similar product out there, I know. But show me one of them where you can failover without becoming the single point of failure? We've just recently seen ... or better "where not able to see anything any more" how failures tend to ripple through systems - half of the US East Coast was dark. So where is the replication system where a slave becomes the "master", and not a standalone server. Show me one that has a clear concept of failback, one that has hot-join as a primary design goal. These are the features that I expect if something is labeled "Enterprise Level". As far as my ideas for multi-master go, it will be a synchronous solution using group communication. My idea is "group commit" instead of 2-Phase ... and an early stage test hack has replicated some update 3 weeks ago. The big challange will be to integrate the two systems so that a node can start as an asynchronous Slony-I slave, catch up ... and switch over to synchronous multimaster without stopping the cluster. I have no clue yet how to do that, but I refuse to think smaller. Jan -- #==# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #== [EMAIL PROTECTED] # ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [HACKERS] Proposal for a cascaded master-slave replication system
Jan, I am wondering if you are familar with the work covered in 'Recovery in Parallel Database Systems' by Svein-Olaf Hvasshovd (Vieweg) ? The book is an excellent detailed description covering high availablility DB implementations. I think your right on by not thinking smaller!! Jordan Henderson On Wednesday 12 November 2003 10:45, Jan Wieck wrote: > Hans-Jürgen Schönig wrote: > > Jan, > > > > First of all we really appreciate that this is going to be an Open > > Source project. > > There is something I wanted to add from a marketing point of view: I > > have done many public talks in the 2 years or so. There is one question > > people keep asking me: "How about the pgreplication project?". In every > > training course, at any conference people keep asking for synchronous > > replication. We have offered this people some async solutions which are > > already out there but nobody seems to be interested in having it (my > > person impression). People keep asking for a sync approach via email but > > nobody seems to care about an async approach. This does not mean that > > async is bad but we can see a strong demand for synchronous replication. > > > > Meanwhile we seem to be in a situation where PostgreSQL is rather > > competing against Oracle than against MySQL. In our case there are more > > people asking for Oracle -> Pg migration than for MySQL -> Pg. MySQL > > does not seem to be the great enemy because most people know that it is > > an inferior product anyway. What I want to point out is that some people > > want an alternative Oracle's Real Application Cluster. They want load > > balancing and hot failover. Even data centers asking for replication did > > not want to have an async approach in the past. > > Hans-Jürgen, > > we are well aware of the high demand for multi-master replication > addressing load balancing and clustering. We have that need ourself as > well and I plan to work on a follow-up project as soon as Slony-I is > released. But as of now, we see a higher priority for a reliable master > slave system that includes the cascading and backup features described > in my concept. There are a couple of different similar product out > there, I know. But show me one of them where you can failover without > becoming the single point of failure? We've just recently seen ... or > better "where not able to see anything any more" how failures tend to > ripple through systems - half of the US East Coast was dark. So where is > the replication system where a slave becomes the "master", and not a > standalone server. Show me one that has a clear concept of failback, one > that has hot-join as a primary design goal. These are the features that > I expect if something is labeled "Enterprise Level". > > As far as my ideas for multi-master go, it will be a synchronous > solution using group communication. My idea is "group commit" instead of > 2-Phase ... and an early stage test hack has replicated some update 3 > weeks ago. The big challange will be to integrate the two systems so > that a node can start as an asynchronous Slony-I slave, catch up ... and > switch over to synchronous multimaster without stopping the cluster. I > have no clue yet how to do that, but I refuse to think smaller. > > > Jan ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
Re: [HACKERS] Proposal for a cascaded master-slave replication system
Jan Wieck wrote: Dear community, for some reason the post I sent yesterday night still did not show up on the mailing lists. I have set up some links on the developers side under http://developer.postgresql.org/~wieck/slony1.html The concept will be the base for some of my work as a Software Engineer here at Afilias USA INC. in the near future. Afilias is like many of you in need of reliable and performant replication solutions for backup and failover purposes. We started this work a couple of weeks ago by defining the goals and required features for our usage of PostgreSQL. Slony-I will be the first of 2 distinct replication systems designed with the 24/7 datacenter in mind. We want to build this system as a community project. The plan was from the beginning to release the product under the BSD license. And we think it is best to start it as such and to ask for suggestions during the design phase already. I would like to start developing the replication engine itself as soon as possible. And as a PostgreSQL CORE developer I will sure put some of my spare time into this as well. On the other hand there is absolutely no design other than "they mostly call some stored procedures" done for the frontend tools yet, and I think that we need some real good admin tools in the end. I look forward to your comments. Jan Jan, First of all we really appreciate that this is going to be an Open Source project. There is something I wanted to add from a marketing point of view: I have done many public talks in the 2 years or so. There is one question people keep asking me: "How about the pgreplication project?". In every training course, at any conference people keep asking for synchronous replication. We have offered this people some async solutions which are already out there but nobody seems to be interested in having it (my person impression). People keep asking for a sync approach via email but nobody seems to care about an async approach. This does not mean that async is bad but we can see a strong demand for synchronous replication. Meanwhile we seem to be in a situation where PostgreSQL is rather competing against Oracle than against MySQL. In our case there are more people asking for Oracle -> Pg migration than for MySQL -> Pg. MySQL does not seem to be the great enemy because most people know that it is an inferior product anyway. What I want to point out is that some people want an alternative Oracle's Real Application Cluster. They want load balancing and hot failover. Even data centers asking for replication did not want to have an async approach in the past. I just wanted to mention that because personally I don't have the impression that an additional async project is worth the effort. Note: This does not mean that it is bad to have one more product ;). Cheers, Hans -- Cybertec Geschwinde u Schoenig Ludo-Hartmannplatz 1/14, A-1160 Vienna, Austria Tel: +43/2952/30706 or +43/660/816 40 77 www.cybertec.at, www.postgresql.at, kernel.cybertec.at ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [HACKERS] Proposal for a cascaded master-slave replication system
Jan, This is EXACTLY what we have been waiting for (years) :) :) :). If you need somebody for testing or documentation just drop me a line. Cheers, Hans Jan Wieck wrote: Hans-Jürgen Schönig wrote: Jan, First of all we really appreciate that this is going to be an Open Source project. There is something I wanted to add from a marketing point of view: I have done many public talks in the 2 years or so. There is one question people keep asking me: "How about the pgreplication project?". In every training course, at any conference people keep asking for synchronous replication. We have offered this people some async solutions which are already out there but nobody seems to be interested in having it (my person impression). People keep asking for a sync approach via email but nobody seems to care about an async approach. This does not mean that async is bad but we can see a strong demand for synchronous replication. Meanwhile we seem to be in a situation where PostgreSQL is rather competing against Oracle than against MySQL. In our case there are more people asking for Oracle -> Pg migration than for MySQL -> Pg. MySQL does not seem to be the great enemy because most people know that it is an inferior product anyway. What I want to point out is that some people want an alternative Oracle's Real Application Cluster. They want load balancing and hot failover. Even data centers asking for replication did not want to have an async approach in the past. Hans-Jürgen, we are well aware of the high demand for multi-master replication addressing load balancing and clustering. We have that need ourself as well and I plan to work on a follow-up project as soon as Slony-I is released. But as of now, we see a higher priority for a reliable master slave system that includes the cascading and backup features described in my concept. There are a couple of different similar product out there, I know. But show me one of them where you can failover without becoming the single point of failure? We've just recently seen ... or better "where not able to see anything any more" how failures tend to ripple through systems - half of the US East Coast was dark. So where is the replication system where a slave becomes the "master", and not a standalone server. Show me one that has a clear concept of failback, one that has hot-join as a primary design goal. These are the features that I expect if something is labeled "Enterprise Level". As far as my ideas for multi-master go, it will be a synchronous solution using group communication. My idea is "group commit" instead of 2-Phase ... and an early stage test hack has replicated some update 3 weeks ago. The big challange will be to integrate the two systems so that a node can start as an asynchronous Slony-I slave, catch up ... and switch over to synchronous multimaster without stopping the cluster. I have no clue yet how to do that, but I refuse to think smaller. Jan -- Cybertec Geschwinde u Schoenig Ludo-Hartmannplatz 1/14, A-1160 Vienna, Austria Tel: +43/2952/30706 or +43/660/816 40 77 www.cybertec.at, www.postgresql.at, kernel.cybertec.at ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [HACKERS] Proposal for a cascaded master-slave replication system
Jordan Henderson wrote: Jan, I am wondering if you are familar with the work covered in 'Recovery in Parallel Database Systems' by Svein-Olaf Hvasshovd (Vieweg) ? The book is an excellent detailed description covering high availablility DB implementations. No, but it sounds like something I allways wanted to have. I think your right on by not thinking smaller!! Thanks Jan Jordan Henderson On Wednesday 12 November 2003 10:45, Jan Wieck wrote: Hans-Jürgen Schönig wrote: > Jan, > > First of all we really appreciate that this is going to be an Open > Source project. > There is something I wanted to add from a marketing point of view: I > have done many public talks in the 2 years or so. There is one question > people keep asking me: "How about the pgreplication project?". In every > training course, at any conference people keep asking for synchronous > replication. We have offered this people some async solutions which are > already out there but nobody seems to be interested in having it (my > person impression). People keep asking for a sync approach via email but > nobody seems to care about an async approach. This does not mean that > async is bad but we can see a strong demand for synchronous replication. > > Meanwhile we seem to be in a situation where PostgreSQL is rather > competing against Oracle than against MySQL. In our case there are more > people asking for Oracle -> Pg migration than for MySQL -> Pg. MySQL > does not seem to be the great enemy because most people know that it is > an inferior product anyway. What I want to point out is that some people > want an alternative Oracle's Real Application Cluster. They want load > balancing and hot failover. Even data centers asking for replication did > not want to have an async approach in the past. Hans-Jürgen, we are well aware of the high demand for multi-master replication addressing load balancing and clustering. We have that need ourself as well and I plan to work on a follow-up project as soon as Slony-I is released. But as of now, we see a higher priority for a reliable master slave system that includes the cascading and backup features described in my concept. There are a couple of different similar product out there, I know. But show me one of them where you can failover without becoming the single point of failure? We've just recently seen ... or better "where not able to see anything any more" how failures tend to ripple through systems - half of the US East Coast was dark. So where is the replication system where a slave becomes the "master", and not a standalone server. Show me one that has a clear concept of failback, one that has hot-join as a primary design goal. These are the features that I expect if something is labeled "Enterprise Level". As far as my ideas for multi-master go, it will be a synchronous solution using group communication. My idea is "group commit" instead of 2-Phase ... and an early stage test hack has replicated some update 3 weeks ago. The big challange will be to integrate the two systems so that a node can start as an asynchronous Slony-I slave, catch up ... and switch over to synchronous multimaster without stopping the cluster. I have no clue yet how to do that, but I refuse to think smaller. Jan ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings -- #==# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #== [EMAIL PROTECTED] # ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [HACKERS] New approach to ye olde cross-datatype indexing problem
I wrote: > What if we attack the problem head on --- make indexes able to handle > cross-datatype comparison operators directly? That would solve the > problem without *any* semantic side-effects. AFAIR we've never seriously > considered doing that; I guess anyone who thought of it dismissed it as > too hard. But I've been studying the idea for the last day or so and I'm > starting to think that it's eminently practical. In fact, it works --- as of CVS tip, you can do: regression=# create table foo (f1 bigint primary key); NOTICE: CREATE TABLE / PRIMARY KEY will create implicit index "foo_pkey" for table "foo" CREATE TABLE regression=# explain select * from foo where f1 = 44; QUERY PLAN Index Scan using foo_pkey on foo (cost=0.00..4.82 rows=1 width=8) Index Cond: (f1 = 44) (2 rows) or even regression=# explain select * from foo where f1 = 44::int2; QUERY PLAN Index Scan using foo_pkey on foo (cost=0.00..4.82 rows=1 width=8) Index Cond: (f1 = 44::smallint) (2 rows) As seems to happen every time I touch the btree code, I learned a few formerly-undocumented details about how it works :-(. It turns out that allowing an operator class to include operators between the indexed datatype and other datatypes isn't really sufficient for everything the backend wants to do. In particular, there are places where the system wants to determine which of two inequalities is a tighter constraint. For example, consider a query like SELECT ... WHERE x > 4 AND x >= 5; The x > 4 condition is redundant and can be thrown away. The btree subsystem actually has code to do this, and in fact was *depending on* the assumption that it could do so, because there were bits of logic that assumed they'd not have to deal with more than one inequality constraint at a time. Now in order to do this, the code has to compare the constants (4 and 5). This is no problem as long as an operator class deals with just a single datatype --- you can pick the needed comparison operator out of the opclass. However, in a cross-type situation you could have something like SELECT ... WHERE int8var > 4::int4 AND int8var >= 5::int2; where the operators appearing in the query are int84lt and int82le. To determine which condition is a tighter constraint, you'd need to apply an int4-vs-int2 comparison --- and there are no such operators available in the int8 opclass. Since the whole point of an opclass is to encapsulate a particular comparison semantics, it's not reasonable for the system to guess at which comparison operator on the other type(s) is compatible with the opclass it's working with. And coercing the comparison values to the opclass' own datatype is no good (that opens up all the problems we wanted to avoid, such as possible overflow). For the moment, I've just punted this problem; the btree code will not optimize away redundant cross-type comparisons. (I went and fixed the places that failed to cope with extra conditions.) It would be nice to do better, but it's surely a much lower priority than getting cross-type comparisons to be indexed at all. Looking to the long term of how we might deal with this if we wanted to invest more effort, I can think of a couple of ways to proceed. One is to allow an opclass to contain operators that aren't directly usable with the index at all, but do apply to datatypes that are used within the opclass --- for instance, we'd have to add all the int4-vs-int4 operators, *and* the int4-vs-int2 and int2-vs-int2 operators, into the int8 opclass; likewise for the int2 and int4 opclasses. This seems like a recipe for bloating pg_amop quite a lot; there'd be a lot of redundancy among these opclasses. The other idea I had was to somehow create associations of opclasses, so that you could say that "int2_ops, int4_ops, and int8_ops all sort values compatibly, so it's okay to go find the operator you need in int4_ops or int2_ops". I don't have any concrete idea about how to represent that though. Any thoughts about that, or other ideas altogether? For the moment I'm satisfied with what we have; this is already a big step forward on a problem that's been there forever. But there's always something else that could be done ... regards, tom lane ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
[HACKERS] ARC buffer strategy committed
I have just committed the implementation of ARC into the 7.5devel tree. This doesn't contain the vacuum page delay, or the test implementation of the background writer based on the checkpoint process. These two need some more discussion and bouncing around. Jan -- #==# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #== [EMAIL PROTECTED] # ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [HACKERS] ARC buffer strategy committed
Jan Wieck <[EMAIL PROTECTED]> writes: > I have just committed the implementation of ARC into the 7.5devel tree. I'm seeing a whole bunch of regression test failures that weren't there half an hour ago ... regards, tom lane ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
Re: [HACKERS] ARC buffer strategy committed
Tom Lane wrote: Jan Wieck <[EMAIL PROTECTED]> writes: I have just committed the implementation of ARC into the 7.5devel tree. I'm seeing a whole bunch of regression test failures that weren't there half an hour ago ... regards, tom lane Oh ... hmmm ... will check Jan -- #==# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #== [EMAIL PROTECTED] # ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [HACKERS] ARC buffer strategy committed
Jan Wieck wrote: Tom Lane wrote: Jan Wieck <[EMAIL PROTECTED]> writes: I have just committed the implementation of ARC into the 7.5devel tree. I'm seeing a whole bunch of regression test failures that weren't there half an hour ago ... regards, tom lane Oh ... hmmm ... will check That's strange, if I reverse apply my patch I have buffer leak's and all kinds of crap. Can't even initdb. Who else applied patches tonight? Jan -- #==# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #== [EMAIL PROTECTED] # ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [HACKERS] 7.4RC2 regression failur and not running stats collector process on Solaris
On Wed, 12 Nov 2003 13:46:52 -0500 Tom Lane <[EMAIL PROTECTED]> wrote: > Kurt Roeckx <[EMAIL PROTECTED]> writes: > > It's always a good idea to put localhost into dns too. > > Yeah, but "localhost" *is* resolving as something on Kiyoshi's > machine, else a different error message would have appeared. > > I'm wondering just what it resolved to though --- maybe we should > have made the error messages more verbose, or added a debug-level > message to show what addresses are being tried. > I tried nslookup on Kiyoshi's machine. $ nslookup localhost Server: name.server.mydomain Address: xxx.xx.xx.xxx : : : (failed test) ^C $ nslookup 127.0.0.1 Server: mail.nagoya2.jrc.or.jp Address: 172.20.12.11 Name:localhost Address: 127.0.0.1 (succesful test) $ /etc/resolv.conf domin mydomain nameserver xxx.xx.xx.xxx /etc/nsswitch.conf hosts: files dns ipnodes:files dns Is it necessary to start a DNS server to bind 'localhost' in Kiyoshi's machine? Reference URL http://sunsolve.sun.com/pub-cgi/retrieve.pl?doc=fsunone/3877 -- Kiyoshi Sawada ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [HACKERS] ARC buffer strategy committed
Jan Wieck <[EMAIL PROTECTED]> writes: > That's strange, if I reverse apply my patch I have buffer leak's and all > kinds of crap. Can't even initdb. Who else applied patches tonight? Well, there was my operator-class patch (which went nowhere near the buffer stuff) and also the initdb-in-C code (new at least to me). But on my HP machine all that stuff seems to play nice. On my Linux machine, which I cvs updated just a little bit later, problems :-( regards, tom lane ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [HACKERS] ARC buffer strategy committed
Jan Wieck <[EMAIL PROTECTED]> writes:
> That's strange, if I reverse apply my patch I have buffer leak's and all
> kinds of crap. Can't even initdb. Who else applied patches tonight?
FWIW, I can initdb --- that seems fine --- but the regression tests spew
lots of small fragments. What I suspect is that the bufmgr is
occasionally returning the wrong buffer :-(
regression.diffs attached. Also, "grep TRAP postmaster.log" shows
this is the cause of the coredump midway through:
TRAP: FailedAssertion("!(( (&(cdb_replace->buf_tag))->rnode.tblNode ==
(&(buf->tag))->rnode.tblNode && (&(cdb_replace->buf_tag))->rnode.relNode ==
(&(buf->tag))->rnode.relNode && (&(cdb_replace->buf_tag))->blockNum ==
(&(buf->tag))->blockNum ))", File: "freelist.c", Line: 610)
Note this is with a serial test (make installcheck). Chris' very
different-looking results may be because he was doing parallel tests;
or there might be something else involved.
regards, tom lane
*** ./expected/name.out Sun May 25 20:11:28 2003
--- ./results/name.out Wed Nov 12 22:16:27 2003
***
*** 90,126
(6 rows)
SELECT '' AS seven, c.f1 FROM NAME_TBL c WHERE c.f1 ~ '.*';
! seven | f1
! ---+-
!| 1234567890ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890ABCDEFGHIJKLMNOPQ
!| 1234567890abcdefghijklmnopqrstuvwxyz1234567890abcdefghijklmnopq
!| asdfghjkl;
!| 343f%2a
!| d34aaasdf
!|
!| 1234567890ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890ABCDEFGHIJKLMNOPQ
! (7 rows)
!
SELECT '' AS zero, c.f1 FROM NAME_TBL c WHERE c.f1 !~ '.*';
! zero | f1
! --+
! (0 rows)
!
SELECT '' AS three, c.f1 FROM NAME_TBL c WHERE c.f1 ~ '[0-9]';
! three | f1
! ---+-
!| 1234567890ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890ABCDEFGHIJKLMNOPQ
!| 1234567890abcdefghijklmnopqrstuvwxyz1234567890abcdefghijklmnopq
!| 343f%2a
!| d34aaasdf
!| 1234567890ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890ABCDEFGHIJKLMNOPQ
! (5 rows)
!
SELECT '' AS two, c.f1 FROM NAME_TBL c WHERE c.f1 ~ '.*asdf.*';
! two | f1
! -+
! | asdfghjkl;
! | d34aaasdf
! (2 rows)
!
DROP TABLE NAME_TBL;
--- 90,104
(6 rows)
SELECT '' AS seven, c.f1 FROM NAME_TBL c WHERE c.f1 ~ '.*';
! ERROR: could not access status of transaction 2752512
! DETAIL: could not open file "/home/tgl/testversion/data/pg_clog/0002": No such file
or directory
SELECT '' AS zero, c.f1 FROM NAME_TBL c WHERE c.f1 !~ '.*';
! ERROR: could not access status of transaction 2752512
! DETAIL: could not open file "/home/tgl/testversion/data/pg_clog/0002": No such file
or directory
SELECT '' AS three, c.f1 FROM NAME_TBL c WHERE c.f1 ~ '[0-9]';
! ERROR: could not access status of transaction 2752512
! DETAIL: could not open file "/home/tgl/testversion/data/pg_clog/0002": No such file
or directory
SELECT '' AS two, c.f1 FROM NAME_TBL c WHERE c.f1 ~ '.*asdf.*';
! ERROR: could not access status of transaction 2752512
! DETAIL: could not open file "/home/tgl/testversion/data/pg_clog/0002": No such file
or directory
DROP TABLE NAME_TBL;
==
*** ./expected/varchar.out Mon May 21 12:54:46 2001
--- ./results/varchar.out Wed Nov 12 22:16:27 2003
***
*** 2,111
-- VARCHAR
--
CREATE TABLE VARCHAR_TBL(f1 varchar(1));
INSERT INTO VARCHAR_TBL (f1) VALUES ('a');
INSERT INTO VARCHAR_TBL (f1) VALUES ('A');
-- any of the following three input formats are acceptable
INSERT INTO VARCHAR_TBL (f1) VALUES ('1');
INSERT INTO VARCHAR_TBL (f1) VALUES (2);
INSERT INTO VARCHAR_TBL (f1) VALUES ('3');
-- zero-length char
INSERT INTO VARCHAR_TBL (f1) VALUES ('');
-- try varchar's of greater than 1 length
INSERT INTO VARCHAR_TBL (f1) VALUES ('cd');
! ERROR: value too long for type character varying(1)
INSERT INTO VARCHAR_TBL (f1) VALUES ('c ');
SELECT '' AS seven, VARCHAR_TBL.*;
! seven | f1
! ---+
!| a
!| A
!| 1
!| 2
!| 3
!|
!| c
! (7 rows)
!
SELECT '' AS six, c.*
FROM VARCHAR_TBL c
WHERE c.f1 <> 'a';
! six | f1
! -+
! | A
! | 1
! | 2
! | 3
! |
! | c
! (6 rows)
!
SELECT '' AS one, c.*
FROM VARCHAR_TBL c
WHERE c.f1 = 'a';
! one | f1
! -+
! | a
! (1 row)
!
SELECT '' AS five, c.*
FROM VARCHAR_TBL c
WHERE c.f1 < 'a';
! five | f1
! --+
! | A
! | 1
! | 2
! | 3
! |
! (5 rows)
!
SELECT '' AS six, c.*
FROM VARCHAR_TBL c
WHERE c.f1 <= 'a';
! six | f1
! -+
! | a
! | A
! | 1
! | 2
!
Re: [HACKERS] ARC buffer strategy committed
Jan Wieck <[EMAIL PROTECTED]> writes: > Can't even initdb. Who else applied patches tonight? FWIW, I did "make distclean", full rebuild, initdb on both HP and Linux machines. The HP passes both serial and parallel regression tests. The Linux machine initdb's, and simple manual queries seem to work, but regression tests fail hard. "diff -r" confirms that the only difference in the source trees is your ARC patch. Configuration in both cases is --enable-cassert --enable-debug ... regards, tom lane ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [HACKERS] Successful proting report for 7.4beta4 on Irix and Altix
Robert E. Bruccoleri wrote: > Dear All, > I have compiled and tested Postgresql-7.4beta4 on SGI Irix > 6.5.20 using the MIPS Pro 7.4 compilers and on SGI Altix using the gcc > 2.96 compilers and both ports passed all tests within the limits of > numerical accuracy. > What is the URL for reporting successful ports? --Bob Right here. I just need uname -a from both machines. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [HACKERS] rpm support for 7.4 and beyond
On Wednesday 12 November 2003 08:57 am, Robert Treat wrote: > I haven't seen any discussion on the topic, so thought I might start > one. We currently provide rpms for Red Hat 7.x, 8.x, and 9, and I'm > assuming that we'll have rpms for all of those for 7.4 once release time > rolls around. However, given that Red Hat is dropping support for 7.x > and 8.x version on Jan 1st, and Red Hat 9 next April, do the current rpm > builders forsee any issues with providing rpms for those platforms in > the future? The biggest issue is going to be 'will it build' on those releases. The tcl version deal (with tcl prior to 8.1) will prevent building on Red Hat 6.2 (which we have tried to support, to varying degrees). I am following the Fedora Legacy project; they purpose to support as a cummunity those older RHL releases. My hands are somewhat tied at the present to only supporting what I actively run. That is currently RHL 8.0 and Fedora Core 1. (not 1.0, incidentally; there is no minor version). I am migrating to Fedora Core 1 across the board, and will be migrating to FC 2 when it comes out (I need the 2.6 kernel's features). While I could easily install and build up a buildfarm for the others, if I'm not actively using it bitrot is more likely to set in. But as part of Fedora Legacy I may be doing this anyway. Depends on the business case I can make for it at PARI. Frankly, supporting the older stuff is a real pain, primarily because people expect an rpm -Uvh to work in a sane manner, which it currently does not do. That, BTW, is why there haven't been any RPMs of 7.4 yet: I plan on having a migration strategy in place first, although if I can't get it working right by final I may just provide the same sort of RPM as prior. The idea is to have 7.2.4 and 7.3.[45] RPMs with the migration strategy built in concurrently to the 7.4 RPMs -- this way, the user can update to the same major version first, then _install_ (NOT upgrade) the 7.4 RPM on top. This works with the linux kernel now; it shouldn't be too hard to make it work with postgresql. > Also (and maybe someone from Red Hat can weigh in here) are there any > plans from Red Hat to release RHEL rpms for postgresql in the future, > and/or plans to make sure the community rpm builders would have access > to those platforms in order to build rpms against them? The RHEL3 beta, taroon, had the rh-postgresql rpms, including the server packages. I am also intrigued by the community-driven rhel-rebuild and cAosity projects. If those stay close to the 'official' thing from Red Hat, then a postgresql rpm built there should run on real RHEL. Having said that, I plan on budgeting for a copy of RHEL WS. I can build there. I have had pretty good results with my volunteer team thus far; I hope that continues. What I do want to ask for, though, is a little patience if the 7.4 release happens with no RPMs concurrently released. I do my best; but I have a job to do. That job is currently very busy. -- Lamar Owen Director of Information Technology Pisgah Astronomical Research Institute 1 PARI Drive Rosman, NC 28772 (828)862-5554 www.pari.edu ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
Re: [HACKERS] rpm support for 7.4 and beyond
On Wed, 12 Nov 2003, Lamar Owen wrote: > The biggest issue is going to be 'will it build' on those releases. > The tcl version deal (with tcl prior to 8.1) Tom applied a patch so that the build will continue to work on 8.0.x ... or is this some other issue? ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
Re: [HACKERS] 7.4RC2 regression failur and not running stats collector process on Solaris
On Thu, 13 Nov 2003 11:39:49 +0900 Kiyoshi Sawada <[EMAIL PROTECTED]> wrote: > $ nslookup localhost > Server: name.server.mydomain > Address: xxx.xx.xx.xxx > : : : > (failed test) > Is it necessary to start a DNS server to bind 'localhost' in Kiyoshi's machine? > I got bind-9.2.2-sol8-intel-local pkg from sun freewear and install to /usr/local. /usr/local/bin/nslookup(ISC-nslookup) was tried on the state where /usr/local/bin/bind(ISC-bind) is not started yet. $ /usr/local/bin/nslookup localhost Note: nslookup is deprecated and may be removed from future releases. Consider using the `dig' or `host' programs instead. Run nslookup with the `-sil[ent]' option to prevent this message from appearing. Server: xxx.xx.xx.xxx Address:xxx.xx.xx.xxx#53 Name: localhost Address: 127.0.0.1 (succesful test) $ /usr/local/bin/nslookup 127.0.0.1 Note: nslookup is deprecated and may be removed from future releases. Consider using the `dig' or `host' programs instead. Run nslookup with the `-sil[ent]' option to prevent this message from appearing. Server: xxx.xx.xx.xxx Address:xxx.xx.xx.xxx#53 1.0.0.127.in-addr.arpa name = localhost. (succesful test) -- Kiyoshi Sawada ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [HACKERS] rpm support for 7.4 and beyond
On Wed, 2003-11-12 at 23:44, Lamar Owen wrote: > My hands are somewhat tied at the present to only supporting what I actively > run. That is currently RHL 8.0 and Fedora Core 1. (not 1.0, incidentally; > there is no minor version). Have you tried mach? http://thomas.apestaart.org/projects/mach/ I know it is used successfully by several 3rd party rpm packaging sites such as freshrpms.net. Also, it's very easy to setup. >From the release notes: WHAT IS IT -- mach allows you to set up clean roots from scratch for any distribution or distribution variation supported. This clean build root can be used to run jailed services, create disk images, or build clean packages. mach can currently set up roots for the following distributions: - Red Hat 7.0 (basic, updated, FreshRPMS) - Red Hat 7.1 (basic, updated, FreshRPMS) - Red Hat 7.2 (basic, updated, FreshRPMS, JPackage) - Red Hat 7.3 (basic, updated, FreshRPMS, JPackage) - Red Hat 8.0 (basic, updated, Fedora, JPackage, GStreamer, FreshRPMS) - Red Hat 9 (basic, updated, Fedora, JPackage, GStreamer, FreshRPMS) - Fedora Core 0.94 (basic, updated) - Fedora Core 0.95 (basic, updated) - SuSE 8.1/8.2 - Yellowdog Linux 2.3 (basic, updated, FreshRPMS) - Yellowdog Linux 3.0 (basic, updated, FreshRPMS) - Dave/Dina oven/fridge ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [HACKERS] ARC buffer strategy committed
Tom Lane wrote: Jan Wieck <[EMAIL PROTECTED]> writes: That's strange, if I reverse apply my patch I have buffer leak's and all kinds of crap. Can't even initdb. Who else applied patches tonight? FWIW, I can initdb --- that seems fine --- but the regression tests spew lots of small fragments. What I suspect is that the bufmgr is occasionally returning the wrong buffer :-( It does, somehow and magically ... I have not found the exact reason yet. I have backed out my patch for the moment, more tomorrow. Sorry, Jan -- #==# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #== [EMAIL PROTECTED] # ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [GENERAL] [HACKERS] Proposal for a cascaded master-slave replication system
On Wed, Nov 12, 2003 at 02:08:23PM +0100, Hans-J?rgen Sch?nig wrote: > an inferior product anyway. What I want to point out is that some people > want an alternative Oracle's Real Application Cluster. They want load > balancing and hot failover. Even data centers asking for replication did > not want to have an async approach in the past. I think Jan has already outlined his more-distant-future idea, but I'd also like to know whether the people who are asking for a replacement for RAC are willing to invest in it? You could buy some _awfully_ good development time for even a year's worth of licensing for RAC. I get the impression from the Postgres-R list that their biggest obstacle is development resources. People often like to say they need hot-fail-capable, five nines, 24/7/365 systems. For most applications, I just do not believe that, and the truth is that the cost of getting from three nines to four (never mind five) is so great that people cheat: one paragraph has the "five nines" clause, and the next paragraph talks about scheduled downtime. In a real "five nines" system (the phone company, say, or the air traffic control system), the time for scheduled downtime is just the cumulative possible outage at any node when it is being switched with its replacement. Five minutes a year is a pretty high bar to jump, and most people long ago concluded that you don't actually need it for most applications. A -- Andrew Sullivan 204-4141 Yonge Street Afilias CanadaToronto, Ontario Canada <[EMAIL PROTECTED]> M2P 2A8 +1 416 646 3304 x110 ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [HACKERS] Proposal for a cascaded master-slave replication system
Hans-Jürgen Schönig wrote: > Meanwhile we seem to be in a situation where PostgreSQL is rather > competing against Oracle than against MySQL. In our case there are more > people asking for Oracle -> Pg migration than for MySQL -> Pg. MySQL > does not seem to be the great enemy because most people know that it is > an inferior product anyway. I can confirm Hans' impressions --- I get very few questions about MySQL vs. PostgreSQL, at least in the past few years. People still using MySQL at this point know they are using something inferior to PostgreSQL, and if they didn't, the new MySQL licensing has made it abundantly clear. MySQL just isn't in the same league, and probably will never be. What people want is Informix/Oracle/MS-SQL => PostgreSQL. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
