-
Notifications
You must be signed in to change notification settings - Fork 86
Catchup command implementation #392
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
…eral TODO added
…UP_SSH_REMOTE set to "ON"
…ce InstanceState * with simple directory string
…er compilation units
…lsn into wait_wal_and_calculate_stop_lsn()
…position2' into release_2_5-pgpro-4850
…ro/pg_probackup into release_2_5-pgpro-4850
…ro/pg_probackup into release_2_5-pgpro-4850
creates an incremental copy for pages that have changed | ||
since the destination database was shut down cleanly. | ||
For this mode, the destination directory must contain a previous | ||
copy of the database that was shut down cleanly. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
copy of the database that was shut down cleanly.
А зачем именно cleanly
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Из изначальной постановки задачи в PGPRO-4447 и пункт 4 раздела Assumptions в issue #277
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Я считаю, что такое требование сильно уменьшает юзабельность и не дает никакого профита.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Т.е. мы таким образом защищаемся от того, чтобы нам не подсунили destination, который, например, только наполовину восстановлен из бэкапа, но по-моему игра не стоит свеч, проще защититься от этого документацией.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Это надо сделать в этой версии или оставим это ослабление требований на следующую версию?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
И, обрати внимание, мы ещё проверяем на наличие backup_label
Это хорошо, sanity чеков много не бывает. Но на 100% всё равно не защищает: вдруг pg_control успели донести, backup_label - нет.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Так а какие проверки в таком случае будут нормальны/достаточны?
Пока то, о чём ты говоришь -- это просто выпилить одну проверку и один тест. Но уверенности в том, что в итоге будет получен работающий инстанс базы, будет меньше
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Так а какие проверки в таком случае будут нормальны/достаточны?
Ну по логике вещей, все остальные имеющиеся минус эта =)
Но уверенности в том, что в итоге будет получен работающий инстанс базы, будет меньше
Я бы не сказал
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
И к тому же, сейчас комбинация двух проверок (на backup_label и на состояние pg_control) как раз защищает больше.
А вот если проверку pg_control выпилить, то как раз твой сценарий и не сможем поймать.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Мой сценарий (а так же много каких други) мы и так не поймаем, поэтому мы не сильно много проигрываем в плане надежности, но сильно выигрываем в юзабельности.
doc/pgprobackup.xml
Outdated
</listitem> | ||
<listitem> | ||
<para> | ||
<literal>PTRACK</literal> — creates an incremental PTRACK backup tracking |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Звучит как будто тут создается бэкап.
Должно же быть что-то вроде ".. read and copy pnly pages that changes since the point of divergence"
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Такой текст пойдет?
PTRACK — tracking page changes on the fly,
only copies pages that have changed since the point of divergence
of the source and destination databases...
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Звучит норм!
Такой текст пойдет?
PTRACK — tracking page changes on the fly,
only copies pages that have changed since the point of divergence
of the source and destination databases...
Мне кажется, очень важно добавить, что и читаются тоже только изменившие страницы
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good enough for release candidate
<application>pg_probackup</application> can create a copy of a <productname>PostgreSQL</productname> | ||
instance directly, without using the backup catalog. This allows you | ||
to add a new standby server in a parallel mode or to have a standby | ||
server that has fallen behind <quote>catch up</quote> with master. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@indrups
Мой рунглиш, конечно, могуч, но может такая формуливка будет звучать лучше ... or to have a fallen-behind standby server to catch with master
?
No description provided.