Thread: How to know if a database has changed
The installation I'm planning will manage several databases, but not all of them will change every day. In order to planning/scripting the pg_dump usage, I would need to know which databases had some change activity at the end of some day. How can it be done? TIA
On Mon, Dec 11, 2017 at 01:48:44PM -0300, marcelo wrote: > The installation I'm planning will manage several databases, but not all of > them will change every day. > In order to planning/scripting the pg_dump usage, I would need to know which > databases had some change activity at the end of some day. > How can it be done? Implement INSERT/UPDATE/DELETE/TRUNCATE triggers on all tables and DDL triggers on CREATE, ALTER, DROP, SECURITY LABEL, COMMENT, GRANT or REVOKE. Karsten -- GPG key ID E4071346 @ eu.pool.sks-keyservers.net E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346
Hi, a) create triggers on some tables (on update, on delete ...). let the triggers write some log, check the log or let the triggers update one central table adding database, tablenames with "table ... in database ... has been updated, modified' b) enable detailed loggin and check the system log. c) both ;) Regards walter Am 11.12.2017 um 17:48 schrieb marcelo: > The installation I'm planning will manage several databases, but not > all of them will change every day. > In order to planning/scripting the pg_dump usage, I would need to know > which databases had some change activity at the end of some day. > How can it be done? > TIA >
The two non elegant ways I can think of is checking the modification time on the files representing the database and a query that checks the pk of all tables. If they're ordered pk you could store the max of them and then compare, otherwise the max of an updated at column would work as well.
Jim
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
Jim
On December 11, 2017 11:48:44 AM EST, marcelo <marcelo.nicolet@gmail.com> wrote:
The installation I'm planning will manage several databases, but not all
of them will change every day.
In order to planning/scripting the pg_dump usage, I would need to know
which databases had some change activity at the end of some day.
How can it be done?
TIA
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
On 11 December 2017 at 11:48, marcelo <marcelo.nicolet@gmail.com> wrote: > The installation I'm planning will manage several databases, but not all of > them will change every day. > In order to planning/scripting the pg_dump usage, I would need to know which > databases had some change activity at the end of some day. > How can it be done? Hi, Just keep a snapshot of pg_stat_database view (https://www.postgresql.org/docs/10/static/monitoring-stats.html#PG-STAT-DATABASE-VIEW) At the end of the day compare the counters of tup_inserted/tup_updated/tup_deleted or just xact_commit/xact_rollback if the snapshot you kept is different from the current values there were modifications on the database. Maybe there are some caveats, for example CREATE commands are included because they insert data in table catalogs? i don't know, is up to you to check that -- Jaime Casanova www.2ndQuadrant.com PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
Am 11.12.2017 um 17:48 schrieb marcelo: > The installation I'm planning will manage several databases, but not > all of them will change every day. > In order to planning/scripting the pg_dump usage, I would need to know > which databases had some change activity at the end of some day. > How can it be done? > TIA > I think you can use logical replication. Take a view at https://github.com/eulerto/wal2json (for instance, there are other solutions available), just capture the wal and check for changes for every database. it's just a rough idea... Regards, Andreas -- 2ndQuadrant - The PostgreSQL Support Company. www.2ndQuadrant.com
Am 11.12.2017 um 18:26 schrieb Andreas Kretschmer: > it's just a rough idea... ... and not perfect, because you can't capture ddl in this way. Regards, Andreas -- 2ndQuadrant - The PostgreSQL Support Company. www.2ndQuadrant.com
I think there's a more useful question, which is why do you want to do this? If it is just about conditional backups, surely the cost of backup storage is low enough, even in S3 or the like, that a duplicate backup is an afterthought from a cost perspective? Before you start jumping through hoops to make your backups conditional, I'd first do some analysis and figure out what the real cost of the thing I'm trying to avoid actually is, since my guess is that you are deep into a premature optimization here, where either the cost of the duplicate backup isn't consequential or the frequency of duplicate backups is effectively 0. It would always be possible to run some kind of checksum on the backup and skip storing it if it matches the previous backup's checksum if you decide that there truly is value in conditionally backing up the db. Sure, that would result in dumping a db that doesn't need to be dumped, but if your write transaction rate is so low that backups end up being duplicates on a regular basis, then surely you can afford the cost of a pg_dump without any significant impact on performance?
On Mon, Dec 11, 2017 at 10:49 AM, Andreas Kretschmer <andreas@a-kretschmer.de> wrote:
Am 11.12.2017 um 18:26 schrieb Andreas Kretschmer:it's just a rough idea...
... and not perfect, because you can't capture ddl in this way.
Hi Sam
You are right, and here are the reason behind my question: The server where postgres will be installed is not on 24/7. It turns on in the morning and goes off at the end of the day. The idea is that, as part of the shutdown process, a local backup is made. The next day, that backup will be copied to the cloud.
In order not to lengthen the shutdown process, we are trying to limit pg_dump to the databases that have had some change, not so much in their schema as in their data.
Of course, to add a trigger for every table and CUD operation on every database is not an option.
You are right, and here are the reason behind my question: The server where postgres will be installed is not on 24/7. It turns on in the morning and goes off at the end of the day. The idea is that, as part of the shutdown process, a local backup is made. The next day, that backup will be copied to the cloud.
In order not to lengthen the shutdown process, we are trying to limit pg_dump to the databases that have had some change, not so much in their schema as in their data.
Of course, to add a trigger for every table and CUD operation on every database is not an option.
On 11/12/17 23:23, Sam Gendler wrote:
I think there's a more useful question, which is why do you want to do this? If it is just about conditional backups, surely the cost of backup storage is low enough, even in S3 or the like, that a duplicate backup is an afterthought from a cost perspective? Before you start jumping through hoops to make your backups conditional, I'd first do some analysis and figure out what the real cost of the thing I'm trying to avoid actually is, since my guess is that you are deep into a premature optimization here, where either the cost of the duplicate backup isn't consequential or the frequency of duplicate backups is effectively 0. It would always be possible to run some kind of checksum on the backup and skip storing it if it matches the previous backup's checksum if you decide that there truly is value in conditionally backing up the db. Sure, that would result in dumping a db that doesn't need to be dumped, but if your write transaction rate is so low that backups end up being duplicates on a regular basis, then surely you can afford the cost of a pg_dump without any significant impact on performance?On Mon, Dec 11, 2017 at 10:49 AM, Andreas Kretschmer <andreas@a-kretschmer.de> wrote:
Am 11.12.2017 um 18:26 schrieb Andreas Kretschmer:it's just a rough idea...
... and not perfect, because you can't capture ddl in this way.
> The next day, that backup will be copied to the cloud. What does this mean? If it is rsync of a local dump to a remote use the directory dump format - disable compression - then each table which didn't change will 'copy' almost instantly. -- Meetings Coordinator, Michigan Association of Railroad Passengers 537 Shirley St NE Grand Rapids, MI 49503-1754 Phone: 616.581.8010 E-mail: awilliam@whitemice.org GPG#D95ED383 Web: http://www.marp.org
On Tue, Dec 12, 2017 at 07:40:46AM -0500, Adam Tauno Williams wrote: > > The next day, that backup will be copied to the cloud. > > What does this mean? If it is rsync of a local dump to a remote use > the directory dump format - disable compression - then each table which > didn't change will 'copy' almost instantly. I think OP wants to minimize the time spent during shutdown, not during the next day. Karsten -- GPG key ID E4071346 @ eu.pool.sks-keyservers.net E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346
Em 12/12/2017 10:14, marcelo escreveu:
Hi, there are plenty of options for secure and optimized bacup solutions for your scenario.
Since you want to backup to the cloud, why not to use pgBarman with diferential backup + log (wal) shipping?
It is perfect, you will have zero downtime, and you can shutdown your database anytime you want.
In my experience, diferential backups (with rsync and symlinks) provide excellent performance and reduced storage (avoid duplications), and works perfectly well, as it also provides automatic rotation for old backups (you define the rules from a set of options).
I've been using pgBarman since 1.4, and I'm very satisfied with it.
Just my 2c,
Edson Richter
Hi Sam
You are right, and here are the reason behind my question: The server where postgres will be installed is not on 24/7. It turns on in the morning and goes off at the end of the day. The idea is that, as part of the shutdown process, a local backup is made. The next day, that backup will be copied to the cloud.
In order not to lengthen the shutdown process, we are trying to limit pg_dump to the databases that have had some change, not so much in their schema as in their data.
Of course, to add a trigger for every table and CUD operation on every database is not an option.On 11/12/17 23:23, Sam Gendler wrote:I think there's a more useful question, which is why do you want to do this? If it is just about conditional backups, surely the cost of backup storage is low enough, even in S3 or the like, that a duplicate backup is an afterthought from a cost perspective? Before you start jumping through hoops to make your backups conditional, I'd first do some analysis and figure out what the real cost of the thing I'm trying to avoid actually is, since my guess is that you are deep into a premature optimization here, where either the cost of the duplicate backup isn't consequential or the frequency of duplicate backups is effectively 0. It would always be possible to run some kind of checksum on the backup and skip storing it if it matches the previous backup's checksum if you decide that there truly is value in conditionally backing up the db. Sure, that would result in dumping a db that doesn't need to be dumped, but if your write transaction rate is so low that backups end up being duplicates on a regular basis, then surely you can afford the cost of a pg_dump without any significant impact on performance?On Mon, Dec 11, 2017 at 10:49 AM, Andreas Kretschmer <andreas@a-kretschmer.de> wrote:
Am 11.12.2017 um 18:26 schrieb Andreas Kretschmer:it's just a rough idea...
... and not perfect, because you can't capture ddl in this way.
Hi, there are plenty of options for secure and optimized bacup solutions for your scenario.
Since you want to backup to the cloud, why not to use pgBarman with diferential backup + log (wal) shipping?
It is perfect, you will have zero downtime, and you can shutdown your database anytime you want.
In my experience, diferential backups (with rsync and symlinks) provide excellent performance and reduced storage (avoid duplications), and works perfectly well, as it also provides automatic rotation for old backups (you define the rules from a set of options).
I've been using pgBarman since 1.4, and I'm very satisfied with it.
Just my 2c,
Edson Richter