backup | Rsync wrapper script for Mac OS X | Continuous Backup library

be aware I don't have a Windows machine against which to try this, so it's just "best effort"

As best I can tell, your problem is because you are trying to recreate the behavior of win_shell by "manually" invoking that improperly quoted cmd.exe /c business, ending up with cmd.exe /c "cmd.exe /c whatever"; dialing up the ansible verbosity -vv could confirm or deny that pattern

Also, the win_shell docs say to use win_command: unless you have a shell redirect need, which as written your task does not.

I solved my issue by using Docker Compose Wait. Essentially, it adds a wait loop that samples the DB container, and only when it's up, runs migrations and seeds the DB.

My next problem was: those seeds ran every time the container was run - I solved that by instead running a script that runs the seeds, and touchs a semaphore file. If the file exists already, it skips the seeds.

Slice always starts =>(greater than equal) of full snapshot time and <(less than) next full snapshot.

for example 2020-12-14T14:26:20Z [complete: 2020-12-14T14:34:39] for this backup PITR(Slice) is 2020-12-14T14:26:40 - 2020-12-16T17:27:26

if you want to restore then first restore 2020-12-14T14:26:20Z [complete: 2020-12-14T14:34:39] then apply 2020-12-14T14:26:40 - 2020-12-16T17:27:26 this slice and you'll get data till 2020-12-16T17:27:26

You can get more details here https://www.percona.com/doc/percona-backup-mongodb/point-in-time-recovery.html

No. The idea of multiple datafiles supporting a single tablespaces is to be able to use striping. This ofcourse only makes sense if your server has multiple physical storage devices that preferably also have their own io interface.

À table will be in the tablespaces and can allocate space in all available datafiles. So the table data can be in all datafiles.

If your io system does not consist of multiple physical devices you might as well use a bigfile tablespace that just has one big datafile. In older releases this was a restore nightmare because the backup and restore was performed file by file.

You didn't describe omitting pg_replslot during your rsync, as the docs recommend. If you didn't omit it, then now your replica has a replication slot which is a clone of the one on the master. But if nothing ever connects to that slot on the replica and advances the cutoff, then the WAL never gets released to recycling. To fix you just need to shutdown the replica, remove that directory, restart it, (and wait for the next restart point to finish).

Do they need to go to wal_archive folder on the disk just like they go to wal_archive folder on the master node?

No, that is optional not necessary. It is set by archive_mode = always if you want it to happen.

You issue could well be that the database is using WAL (Write-Ahead logging) as opposed to journal mode.

With WAL changes are written to the WAL file (database file name suffixed with -wal and also another file -shm). If the database hasn't been committed and you only backup/restore the database file. You will lose data.

• When fully committed, the -wal file will be 0 bytes or not exist, in which case it is not needed.

From Android 9 the default was changed from journal mode to WAL.

• https://source.android.com/devices/tech/perf/compatibility-wal

Assuming that this is your issue you have some options:-

1. Use Journal mode (e.g. use the SQLiteDatabase disableWriteAheadLogging method)
2. Backup/Restore all 3 files (if they exist)
3. Fully Commit the database and then backup (closing the database should fully commit) and delete/rename the -wal file and -shm file before restoring.

Option 3 would be the recommended way as you then gain the advantages of WAL.

Here's an example of fully checkpointing (a little over the top but it works):-

It's this typo:

try below code:

You cannot specify all the collections to export, but you can specify which collections not to export using --excludeCollection, like this:

mongodump --db=test --excludeCollection=users --excludeCollection=salaries

It is listed on the documentation.