This reverts commit 5a00a041f1.
Implementation is currently not correct. Multiple uploads can share the
same etag but have different paths in the S3 bucket.
A "bad upload" in this context is a upload with a mismatched URL. This can happen when changing the S3 bucket used for uploads and the upload records in the database have not been remapped correctly.
This commit ensures that we reset the `missing_s3_uploads` status count
if there are no inventory files which are at least 2 days older than the
site's restored date.
Otherwise, a site with missing uploads but was subsequntly restored will
be continue to report missing uploads for 2 days.
This commit adds a hidden `s3_inventory_bucket_region` site setting to
specify the region of the `s3_inventory_bucket` when the `S3Inventory`
class initializes an instance of the `S3Helper`. By default, the
`S3Helper` class uses the value of the `s3_region` site setting but the
region of the `s3_inventory_bucket` is not always the same as the
`s3_region` configured.
This commit introduces a hidden `s3_inventory_bucket` site setting which
replaces the `enable_s3_inventory` and `s3_configure_inventory_policy`
site setting.
The reason `enable_s3_inventory` and `s3_configure_inventory_policy`
site settings are removed is because this feature has technically been
broken since it was introduced. When the `enable_s3_inventory` feature
is turned on, the app will because configure a daily inventory policy for the
`s3_upload_bucket` bucket and store the inventories under a prefix in
the bucket. The problem here is that once the inventories are created,
there is nothing cleaning up all these inventories so whoever that has
enabled this feature would have been paying the cost of storing a whole
bunch of inventory files which are never used. Given that we have not
received any complains about inventory files inflating S3 storage costs,
we think that it is very likely that this feature is no longer being
used and we are looking to drop support for this feature in the not too
distance future.
For now, we will still support a hidden `s3_inventory_bucket` site
setting which site administrators can configure via the
`DISCOURSE_S3_INVENTORY_BUCKET` env.
This commit introduces the following changes which allows a site
administrator to mark `Upload` records with the `s3_file_missing`
verification status which will result in the `Upload` record being ignored when
`Discourse.store.list_missing_uploads` is ran on a site where S3 uploads
are enabled and `SiteSetting.enable_s3_inventory` is set to `true`.
1. Introduce `s3_file_missing` to `Upload.verification_statuses`
2. Introduce `Upload.mark_invalid_s3_uploads_as_missing` which updates
`Upload#verification_status` of all `Upload` records from `invalid_etag` to `s3_file_missing`.
3. Introduce `rake uploads:mark_invalid_s3_uploads_as_missing` Rake task
which allows a site administrator to change `Upload` records with
`invalid_etag` verification status to the `s3_file_missing`
verificaton_status.
4. Update `S3Inventory` to ignore `Upload` records with the
`s3_file_missing` verification status.
This commit updates `S3Inventory#files` to ignore S3 inventory files
which have a `last_modified` timestamp which are not at least 2 days
older than `BackupMetadata.last_restore_date` timestamp.
This check was previously only in `Jobs::EnsureS3UploadsExistence` but
`S3Inventory` can also be used via Rake tasks so this protection needs
to be in `S3Inventory` and not in the scheduled job.
In some setups the keys start with "original/" and "optimized/" and in some setups the key is something like "foo/original/", so lets make the filter less strict.
The UPDATE statement could lock the `uploads` table for a very long time
when the `verification_status` of lots of uploads changed. Splitting up
and simplifying the UPDATE solves that problem.
Also, this change ensures that only the needed data from the inventory
gets inserted into the `TEMP TABLE`. For example, there's no need to
have records for optimized images in that table when the `uploads` table
gets updated.
* File.exists? is deprecated and removed in Ruby 3.2 in favor of
File.exist?
* Dir.exists? is deprecated and removed in Ruby 3.2 in favor of
Dir.exist?
Inventory on S3 always lagged, over the past few weeks we are noticing that
1 day of lag is not enough.
We are increasing this to 2, to ensure that we do not get false positive
reports.
When we run the S3 inventory, mark uploads that exist as verified true, those that don't as verified false, and uploads not included in the check / not yet checked as verified nil.
Previously we considered 'upload rows without etags' to be exempt from the check. This is bad, because older/migrated sites might not have etags on all their uploads. We should consider rows without etags to be broken, since we can't check them against the inventory.
This also removes the `by_users` scope. We need all uploads to be working, even ones created by the system user.
This is a very expensive process, and it should only be required in exceptional circumstances. It is possible to run a similar recovery using `rake uploads:recover` (5284d41a8e/lib/upload_recovery.rb (L135-L184))
When we change upload's sha1 (e.g. when resizing images) it won't match the data in the most recent S3 inventory index. With this change the uploads that have been updated since the inventory has been generated are ignored.
In sidekiq, jobs are run in multiple threads within the same process. `cd` affects the entire process, so can cause unexpected issues in other running jobs.
