Use Google Cloud Storage Instead of S3

AI-generated summary

The discussion revolves around using Google Cloud Storage (GCS) instead of Amazon S3 for storing Discourse backups. yyhmsg proposes modifying the AWS endpoint configuration in app.yml to use GCS’s compatibility mode. However, they ended up using a bash script to copy and remove backups manually.

downey agrees that having more options for cloud storage would be beneficial. tuanpembual shares their success in using GCS for backups by adding a special policy and using the default configuration with DISCOURSE_S3_BACKUP_BUCKET and DISCOURSE_BACKUP_LOCATION.

Mark_Schmucker suggests using Cloud Storage Transfer Service to sync backups from S3 to GCS, which doesn’t require programming but may increase storage costs. This approach allows for keeping multiple backups, with a minimal cost of $3/month for 10 GB of data.

I saw another post about this from '14 - but no answer and it seemed like they were making this harder than it should be…

I would like some pointers to see if there is an AWS endpoint configuration within app.yml which could be changed to allow use of google cloud storage instead of s3 - GCS has a compatability mode where all - as I understand it - would need to happen is to change the AWS endpoint:

In your existing tools or libraries, make the following changes:
Change the request endpoint to use the Google Cloud Storage request endpoint.
Replace the Amazon Web Services (AWS) access and secret key with the corresponding Google Cloud Storage access key and secret key (collectively called your Google developer key).

This could be added as an option under /admin/site_settings/category/files and option for AWS endpoint along with the access keys? or can it be done in the app.yml? [AWS_ENDPOINT = “/xxxx/xxx”] already?

Thanks!

2 Me gusta

I ended up doing this, just to get it working:

/bin/bash
 This script will copy the latest discourse backup to google cloud
 Then it will remove the previous backup from google cloud

LATESTBACKUP=`ls /var/discourse/shared/standalone/backups/default/*.gz -t1 | head -n1`
PREVBACKUPWDIR=`ls /var/discourse/shared/standalone/backups/default/*.gz -1tr | head -1`
PREVBACKUP=`find $PREVBACKUPWDIR -exec basename {} \;`

/home/user/gsutil/gsutil cp $LATESTBACKUP gs://yyh-community-backups
/home/user/gsutil/gsutil rm gs://your-backup-bucket/$PREVBACKUP

Which works, but I still think greater use of google cloud storage could be used with the s3 compatibility mode as described above. But maybe I’m the only one who wants to use non amazon tools.

3 Me gusta

It’d be great to have a few options “out of the box”!

Hola

Estoy probando la copia de seguridad en GCS usando la configuración predeterminada.

DISCOURSE_S3_BACKUP_BUCKET: falcoland-files/backup
DISCOURSE_BACKUP_LOCATION: s3

y agregué una política especial en GCS

policy = {
    "serviceAccount:${google_service_account.discourse.email}" = [
      "roles/storage.legacyBucketOwner"
    ]
  }

Se envió la copia de seguridad a GCS correctamente.

Espero que esto ayude a alguien como yo que pasó una semana en aplicarlo.

Almacenamiento GCS

6 Me gusta

Esto probablemente roza la paranoia, pero guardo las copias de seguridad de Discourse en AWS S3 (lo cual es compatible de forma nativa) y luego sincronizo esas copias desde S3 a GCS usando:

No se requiere programación. Esto aumentará ligeramente tus costos de almacenamiento, por lo que no ayuda si tu objetivo es ahorrar dinero. Conservo las 10 copias de seguridad más recientes; cada una ocupa 10 GB y me cuesta 3 dólares al mes en Google.