Thread: Audio interview
I did an audio interview today, and it is online now: http://bsdtalk.blogspot.com/2006/02/bsdtalk015-interview-with-postgresql.html -- 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
On Tue, Feb 07, 2006 at 11:43:40PM -0500, Bruce Momjian wrote: > I did an audio interview today, and it is online now: > > http://bsdtalk.blogspot.com/2006/02/bsdtalk015-interview-with-postgresql.html Great interview. You hit a lot of the high points :) You mentioned in-place upgrade scripts. Are those in contrib/ somewhere? On GBorg? On PgFoundry? If not, could you put them somewhere? As far as converting them from shell to Perl, I'm sure you'll find a flock of volunteers to help. Cheers, D -- David Fetter david@fetter.org http://fetter.org/ phone: +1 415 235 3778 Remember to vote!
On Tuesday 07 February 2006 23:43, Bruce Momjian wrote: > I did an audio interview today, and it is online now: > > > http://bsdtalk.blogspot.com/2006/02/bsdtalk015-interview-with-postgresql.html Nice going Bruce. I noticed that both Dru and Dan Langille (creator of fresh ports) have also had interviews done on thissite. cheers
David Fetter wrote: > On Tue, Feb 07, 2006 at 11:43:40PM -0500, Bruce Momjian wrote: > > I did an audio interview today, and it is online now: > > > > http://bsdtalk.blogspot.com/2006/02/bsdtalk015-interview-with-postgresql.html > > Great interview. You hit a lot of the high points :) > > You mentioned in-place upgrade scripts. Are those in contrib/ > somewhere? On GBorg? On PgFoundry? If not, could you put them /contrib/pgupgrade > somewhere? As far as converting them from shell to Perl, I'm sure > you'll find a flock of volunteers to help. Yea, but the problem with modifying the disk pages is still a problem. -- 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
On Wed, Feb 08, 2006 at 09:00:46AM -0500, Bruce Momjian wrote: > David Fetter wrote: > > On Tue, Feb 07, 2006 at 11:43:40PM -0500, Bruce Momjian wrote: > > > > You mentioned in-place upgrade scripts. Are those in contrib/ > > somewhere? On GBorg? On PgFoundry? If not, could you put them > > /contrib/pg_upgrade I see it in the attic, but not in CVS TIP. Is there some way to get it back? Or should it go somewhere else until it's at least slightly working? > > somewhere? As far as converting them from shell to Perl, I'm sure > > you'll find a flock of volunteers to help. > > Yea, but the problem with modifying the disk pages is still a > problem. I understand that not everybody will choose this path, but we've gone to a *lot* of trouble--and as you pointed out, have benefitted directly from the effort--to provide pointy-hair checkboxes like the Windows port. "In-place upgrade" is one of those checkboxes, and I'm pretty confident that getting it working will have at a minimum the same benefits to the rest of the code that making the Windows port did. Cheers, D -- David Fetter david@fetter.org http://fetter.org/ phone: +1 415 235 3778 Remember to vote!
Andrew, > This would be a very fine project for someone to pick up (maybe one of > the corporate supporters could sponsor someone to work on it?) We looked at it for Greenplum but just couldn't justify putting it near the top of the priority list. The work/payoff ratio is terrible. One justification for in-place upgrades is to be faster than dump/reload. However, if we're assuming the possibility of new/modified header fields which could then cause page splits on pages which are 90% capacity, then this time savings would be on the order of no more than 50% of load time, not the 90% of load time required to justify the programming effort involved -- especially when you take into account needing to provide multiple conversions, e.g. 7.3-->8.1, 7.4 --> 8.1, etc. The second reason for in-place upgrade is for large databases where the owner does not have enough disk space for two complete copies of the database. Again, this is not solvable; if we want in-place upgrade to be fault-tolerant, then we need the doubled disk space anyway (you could do a certain amount with compression, but you'd still need 150%-175% space so it's not much help). Overall, it would be both easier and more effective to write a Slony automation wrapper which does the replication, population, and switchover for you. --Josh
On Wed, 2006-02-08 at 11:55 -0800, Josh Berkus wrote: > One justification for in-place upgrades is to be faster than > dump/reload. However, if we're assuming the possibility of new/modified > header fields which could then cause page splits on pages which are 90% > capacity, then this time savings would be on the order of no more than > 50% of load time Well, if you need to start shuffling heap tuples around, you also need to update indexes, in addition to rewriting all the heap pages. This would require work on the order of VACUUM FULL in the worst case, which is pretty expensive. However, we don't change the format of heap or index pages _that_ often. An in-place upgrade script that worked when the heap/index page format has not changed would still be valuable -- only the system catalog format would need to be modified. > The second reason for in-place upgrade is for large databases where the > owner does not have enough disk space for two complete copies of the > database. Again, this is not solvable; if we want in-place upgrade to > be fault-tolerant, then we need the doubled disk space anyway When the heap/index page format hasn't changed, we would only need to backup the system catalogs, which would be far less expensive. -Neil
Josh Berkus <josh@agliodbs.com> writes:
>> This would be a very fine project for someone to pick up (maybe one of
>> the corporate supporters could sponsor someone to work on it?)
> We looked at it for Greenplum but just couldn't justify putting it near
> the top of the priority list. The work/payoff ratio is terrible.
I agree that doing pgupgrade in full generality is probably not worth
the investment required. However, handling the restricted case where
no changes are needed in user tables or indexes would be considerably
easier, and I think it would be worth doing.
If such a tool were available, I don't think it'd be hard to get
consensus on organizing our releases so that it were applicable more
often than not. We could postpone changes that would affect user
table contents until we'd built up a backlog that would all go into
one release. Even a minimal commitment in that line would probably
result in pgupgrade working for at least every other release, and
that would be enough to make it worthwhile if you ask me ...
regards, tom lane
Tom, > If such a tool were available, I don't think it'd be hard to get > consensus on organizing our releases so that it were applicable more > often than not. We could postpone changes that would affect user > table contents until we'd built up a backlog that would all go into > one release. Even a minimal commitment in that line would probably > result in pgupgrade working for at least every other release, and > that would be enough to make it worthwhile if you ask me ... We could even make that our first/second dot difference in the future. That is, 8.2 will be pg-upgradable from 8.1 but 9.0 will not. -- --Josh Josh Berkus Aglio Database Solutions San Francisco
David Fetter wrote: > On Wed, Feb 08, 2006 at 09:00:46AM -0500, Bruce Momjian wrote: > > David Fetter wrote: > > > On Tue, Feb 07, 2006 at 11:43:40PM -0500, Bruce Momjian wrote: > > > > > > You mentioned in-place upgrade scripts. Are those in contrib/ > > > somewhere? On GBorg? On PgFoundry? If not, could you put them > > > > /contrib/pg_upgrade > > I see it in the attic, but not in CVS TIP. Is there some way to get > it back? Or should it go somewhere else until it's at least slightly > working? I think from cvsweb you can get to the Attic files. -- 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
David Fetter wrote: >On Wed, Feb 08, 2006 at 09:00:46AM -0500, Bruce Momjian wrote: > > >>David Fetter wrote: >> >> >>>On Tue, Feb 07, 2006 at 11:43:40PM -0500, Bruce Momjian wrote: >>> >>>You mentioned in-place upgrade scripts. Are those in contrib/ >>>somewhere? On GBorg? On PgFoundry? If not, could you put them >>> >>> >> /contrib/pg_upgrade >> >> > >I see it in the attic, but not in CVS TIP. Is there some way to get >it back? Or should it go somewhere else until it's at least slightly >working? > > There is a pgfoundry project, but it appears to be dead: http://pgfoundry.org/projects/pgupgrade This would be a very fine project for someone to pick up (maybe one of the corporate supporters could sponsor someone to work on it?) cheers andrew
On Feb 8, 2006, at 7:00 AM, Bruce Momjian wrote: > David Fetter wrote: >> On Tue, Feb 07, 2006 at 11:43:40PM -0500, Bruce Momjian wrote: >>> I did an audio interview today, and it is online now: >>> >>> http://bsdtalk.blogspot.com/2006/02/bsdtalk015-interview-with- >>> postgresql.html >> >> Great interview. You hit a lot of the high points :) >> >> You mentioned in-place upgrade scripts. Are those in contrib/ >> somewhere? On GBorg? On PgFoundry? If not, could you put them > > /contrib/pgupgrade > >> somewhere? As far as converting them from shell to Perl, I'm sure >> you'll find a flock of volunteers to help. > > Yea, but the problem with modifying the disk pages is still a problem. Maybe this is totally crazy, but for those not using slony but are using incremental backup and want to upgrade without doing a time consuming dump / reload (this is not actually a problem for me as my data is not so large that a dump reload is a huge problem) would it be possible to apply pgupgrade to the physical backup before you restore, then also alter each WAL record as it is restored so that it restores all new pages in the new format. Then you could do all the work on a different box and quickly switch over to it after the restore is complete. You could eliminate most of the downtime. Is that even feasible? Not something that would help me now but it might make some people very happy (and maybe someday I will need it as well.) Rick
On Feb 8, 2006, at 12:55 PM, Josh Berkus wrote: > Andrew, > >> This would be a very fine project for someone to pick up (maybe >> one of the corporate supporters could sponsor someone to work on it?) > > We looked at it for Greenplum but just couldn't justify putting it > near the top of the priority list. The work/payoff ratio is terrible. > > One justification for in-place upgrades is to be faster than dump/ > reload. However, if we're assuming the possibility of new/modified > header fields which could then cause page splits on pages which are > 90% capacity, then this time savings would be on the order of no > more than 50% of load time, not the 90% of load time required to > justify the programming effort involved -- especially when you take > into account needing to provide multiple conversions, e.g. 7.3-- > >8.1, 7.4 --> 8.1, etc. I just posted an idea for first upgrading a physical backup of the data directory that you would create when doing "Online backups" and then also altering the the WAL log records as they are applied during recovery. That way the actual load time might still be huge but since it could run in parallel with the running server it would probably eliminate 99% of the downtime. Would that be worth the effort? Also all the heavy lifting could be offloaded to a separate box while your production server just keeps running unaffected. > The second reason for in-place upgrade is for large databases where > the owner does not have enough disk space for two complete copies > of the database. Again, this is not solvable; if we want in-place > upgrade to be fault-tolerant, then we need the doubled disk space > anyway (you could do a certain amount with compression, but you'd > still need 150%-175% space so it's not much help). Yeah, anyone who has so much data that they need this feature but isn't willing to back it up is crazy. Plus disk space is cheap. > Overall, it would be both easier and more effective to write a > Slony automation wrapper which does the replication, population, > and switchover for you. Now that is something that I would actually use. I think that a little bit of automation would greatly enhance the number of users using slony. Rick
Ühel kenal päeval, K, 2006-02-08 kell 15:51, kirjutas Tom Lane: > Josh Berkus <josh@agliodbs.com> writes: > >> This would be a very fine project for someone to pick up (maybe one of > >> the corporate supporters could sponsor someone to work on it?) > > > We looked at it for Greenplum but just couldn't justify putting it near > > the top of the priority list. The work/payoff ratio is terrible. > > I agree that doing pgupgrade in full generality is probably not worth > the investment required. However, handling the restricted case where > no changes are needed in user tables or indexes would be considerably > easier, and I think it would be worth doing. How hard would it be to modify postgres so that it can handle multiple heap page formats ? This could come handy for pgupgrade, but my real interest would be to have several task-specific formats supported even in non-upgrade situations, such as a more compact heap page format for read-only archive/analysis tables. -------------- Hannu
Hannu Krosing wrote:
>
> How hard would it be to modify postgres so that it can handle multiple
> heap page formats ?
>
If the on-disk format is changed to add a feature (rather that for some
performance reason), then that would mean that the feature would have to
be available or not per disk page. Wouldn't that cause problems?
Regards,
Jeff Davis
Jeff Davis wrote: > Hannu Krosing wrote: > > > > How hard would it be to modify postgres so that it can handle multiple > > heap page formats ? > > If the on-disk format is changed to add a feature (rather that for some > performance reason), then that would mean that the feature would have to > be available or not per disk page. Wouldn't that cause problems? Yeah, it would be problematic and difficult to handle. For example in subtransactions it would be a hassle to handle the 7.4 heap page format, maybe impossible without race conditions. -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.