decsis / docker-barman Goto Github PK

When I stop docker barman container, lock files are not cleared

-rw------- 1 lxd  lxd     4 Apr  4 13:43 .postgresql.postgresql.rancher.internal-archive-wal.lock
-rw------- 1 lxd  lxd     4 Apr  4 13:43 .postgresql.postgresql.rancher.internal-cron.lock
-rw------- 1 lxd  lxd     4 Apr  4 13:29 .postgresql.postgresql.rancher.internal-receive-wal.lock
-rw------- 1 lxd  lxd     4 Apr  4 13:43 .postgresql.postgresql.rancher.internal-xlogdb.lock

• last backup size and date for each server
• duration times per server and type (full and wal)
• barman check status (0 or 1) for each server

Maybe using barman hook scripts

Use example docker compose up ,but failed at backup all step.
its always ask root@rec_pg password
still failed after volumen .ssh folder that i can ssh inside the barman container

Regarding #7 if you one is using OverlayFS cron execution fall into the bug described here.

Hi,
We are using Kubernetes, Helm + Docker. Developers dynamically (un-)deploy Postgres images on their own. We have to fully automate the backup therefore.

We have to handle one entry per server within "/etc/barman.d". Any idea how we should handle these entries fully automated?
-Markus

barman deb package installs cron task to /etc/cron.d/barman and barman user also has {{barman cron}} task configured resulting in these log messags:

2017-04-04 11:38:01,371 [323] barman.server INFO: Another cron process is already running on server postgresql.postgresql.rancher.internal. Skipping to the next server
2017-04-04 11:38:01,521 [339] barman.wal_archiver INFO: No xlog segments found from streaming for postgresql.postgresql.rancher.internal.
2017-04-04 11:39:01,576 [348] barman.server INFO: Another cron process is already running on server postgresql.postgresql.rancher.internal. Skipping to the next server
2017-04-04 11:39:01,722 [362] barman.wal_archiver INFO: No xlog segments found from streaming for postgresql.postgresql.rancher.internal.
2017-04-04 11:40:01,786 [367] barman.server INFO: Another cron process is already running on server postgresql.postgresql.rancher.internal. Skipping to the next server
2017-04-04 11:40:01,940 [385] barman.wal_archiver INFO: No xlog segments found from streaming for postgresql.postgresql.rancher.internal.
2017-04-04 11:41:01,965 [390] barman.server INFO: Another cron process is already running on server postgresql.postgresql.rancher.internal. Skipping to the next server
2017-04-04 11:41:02,117 [408] barman.wal_archiver INFO: No xlog segments found from streaming for postgresql.postgresql.rancher.internal.

using a logger (instead of print cmds, to enable/disable during devel), one can easily switch on/off what level of debug/info logs one wants to see, between devel and production 👍

• Last wal file synced
• Wal file size

The goal is to have a scheduled recovery task for desired servers that somehow ensures the backup validity. This recovery task should also generate Prometheus metrics.

Maybe this should be accomplished by a companion image.

Port is currently hardcoded (to 8000)
I want to contribute an improvment to have it configurable (keeping old port as default) through 1 of the following ways:

• (opt) cmd-line args (like --listen-port)
• new (opt) ENV var like BARMAN_EXPORTER_PORT
  • PR (for this variant): #18

any preference @rsilva4 ?
IMO using the ENV is easier and consistent to the current impl. though using cmdline args would be more standard

How to use with a remote database?