Commit Graph

56 Commits

Author SHA1 Message Date
Daniel Waterworth 666536cbd1
DEV: Prefer \A and \z over ^ and $ in regexes (#19936) 2023-01-20 12:52:49 -06:00
David Taylor 6417173082
DEV: Apply syntax_tree formatting to `lib/*` 2023-01-09 12:10:19 +00:00
Gerhard Schlager 62be87d5f3
FIX: Filtering rows of S3 inventory files was too strict (#19153)
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.
2022-11-22 21:41:22 +01:00
Gerhard Schlager a597ef7131
PERF: Speed up S3 inventory updates (#19110)
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.
2022-11-20 21:52:30 +01:00
Peter Zhu c5fd8c42db
DEV: Fix methods removed in Ruby 3.2 (#15459)
* 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?
2022-01-05 18:45:08 +01:00
Sam e0c952290b
FIX: increase inventory lag for s3 to 2 days (#11606)
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.
2020-12-30 16:05:42 +11:00
Penar Musaraj 9f6c4ad71a
FIX: inconsistency in S3 inventory config (#11112)
Ensures it matches S3 inventory config generation in our hosting.
2020-11-05 08:39:40 -05:00
Martin Brennan 80268357e7
DEV: Change upload verified column to be integer (#10643)
Per review https://review.discourse.org/t/dev-add-verified-to-uploads-and-fill-in-s3-inventory-10406/14180

Change the verified column for Upload to a verified_status integer column, to avoid having NULL as a weird implicit status.
2020-09-17 13:35:29 +10:00
Penar Musaraj 06b4ca5dc7
FIX: Mark only uploads as verified/unverified in S3 inventory 2020-09-14 10:21:34 -04:00
David Taylor bd0a7553c4
DEV: Detect when s3 inventory failure is caused by etag difference (#10427) 2020-08-13 09:30:28 +10:00
Martin Brennan b950b3fb3f
DEV: Add verified to uploads and fill in S3 inventory (#10406)
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.
2020-08-11 14:43:51 +10:00
David Taylor 16c65a94f7
PERF: Preload S3 inventory data for multisite clusters 2020-07-29 10:31:55 +01:00
David Taylor ec4024fe6d
FIX: Keep by_users check in S3 inventory
Partial revert of 8515d8fa - the by_users check is ensuring we don't raise errors for fixtures
2020-07-21 17:19:56 +01:00
David Taylor 8515d8fae5
FIX: Improve S3 inventory logic
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.
2020-07-21 15:55:53 +01:00
David Taylor 3d65678a13
DEV: Add timestamp columns to optimized_images table (#10199)
This allows us to filter by created/updated date when comparing to an S3 inventory.
2020-07-14 11:50:33 +01:00
David Taylor 7f2b5a446a
PERF: Remove post_upload recovery in daily EnsureS3UploadsExistence job (#10173)
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))
2020-07-06 16:26:40 +01:00
Sam Saffron 38a30a6e96
DEV: correct regression and correct tests
etag change in 31976ecf was incorrect, revert it

Also correct regression in test suite.
2020-07-06 10:56:19 +10:00
Sam Saffron 31976ecfeb
PERF: only update etag when it changes
Previously when synchronizing upload etags we would update every single one
regardless of change.
2020-07-06 10:40:04 +10:00
Jarek Radosz 73b04976e5
FIX: Use updated_at in the S3 inventory job (#8823)
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.
2020-01-31 11:02:44 +01:00
Vinoth Kannan 3b7f5db5ba
FIX: parallel spec system needs a dedicated upload folder for each worker. (#8547) 2019-12-18 11:21:57 +05:30
Osama Sayegh 68708db721 DEV: `S3Inventory#unsorted_files` should always return an array (#8034) 2019-08-23 17:59:31 +10:00
Sam Saffron e53a171916 FIX: hold s3 related distributed locks longer
These operations are pretty expensive and can take multiple minutes due to
networking.

Hold distributed mutex for much longer.
2019-08-15 11:48:44 +10:00
Vinoth Kannan 9919ee1900 FIX: remove the tmp inventory files after the s3 uploads check. 2019-08-13 11:52:57 +05:30
Guo Xiang Tan 8a64b0c8e8 Revert "DEV: Remove unused kwarg and properly check for local missing uploads."
This reverts commit 97769f3d02.

The code is confusing but this change is quite risky. Defer for now
until we can look at it properly.
2019-07-29 14:35:34 +08:00
Guo Xiang Tan 97769f3d02 DEV: Remove unused kwarg and properly check for local missing uploads. 2019-07-29 14:21:06 +08:00
Vinoth Kannan 47deb8b3da FIX: use same id for both original & optimized inventories in multisite setup. 2019-07-25 14:16:47 +05:30
Vinoth Kannan ad04ce9f43 FIX: remove post upload record creation inside 'find_missing_uploads' method. 2019-07-19 01:44:08 +05:30
Vinoth Kannan 35d6fff69e PERF: use url instead of file key in temporary inventory table. 2019-06-13 22:03:58 +05:30
David Taylor ed21128ee6 FIX: Do not change directory when decompressing S3 inventory
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.
2019-06-13 17:13:50 +01:00
Vinoth Kannan d74ee9dbce DEV: skip S3 inventory records without correct multisite prefix. 2019-06-08 18:36:06 +05:30
Vinoth Kannan 2941c77abc FIX: skip upload recovery if file not found in s3 2019-05-21 00:06:36 +05:30
Vinoth Kannan 2a7065c505 FIX: skip uploads without etag in s3 inventory check. 2019-05-20 00:09:52 +05:30
Vinoth Kannan 3172172b52 remove unused local variable
ec84c87ddb
2019-05-16 15:39:13 +05:30
Vinoth Kannan ec84c87ddb FIX: skip validation while recovering uploads from s3
TODO: add tests
2019-05-16 15:37:11 +05:30
Vinoth Kannan 40328f055e FIX: retrieve original filename from s3 object's content disposition header 2019-05-16 09:47:22 +05:30
Guo Xiang Tan dd49be27d3 DEV: Fix undefined variable.
Follow up to e8fafbc123.
2019-05-16 11:28:48 +08:00
Vinoth Kannan f5a217be92 Fix typo in condition value. 2019-05-07 17:09:08 +05:30
Vinoth Kannan e8fafbc123 List and restore missing post uploads from S3 inventory. 2019-05-04 01:16:20 +05:30
Vinoth Kannan 73418aaf73 DEV: Add bucket folder path to inventory id 2019-05-02 04:35:35 +05:30
Vinoth Kannan a8f410a9c5
FEATURE: Create new helper method 'Discourse.stats' (#7388) 2019-04-17 12:45:04 +05:30
Vinoth Kannan 35431a8ddb FIX: set missing count in redis even if zero 2019-04-04 20:05:57 +05:30
Vinoth Kannan df6ef856e6 DEV: save missing s3 uploads count in redis 2019-04-04 19:05:57 +05:30
Guo Xiang Tan 243fb8d9ad Fix the build. 2019-03-13 17:39:07 +08:00
Vinoth Kannan da1ff2da2c FIX: Create and consume temp table inside a transaction (#7030)
To prevent access issue in pgbouncer which runs in transaction pooling
2019-02-20 13:52:40 +11:00
Vinoth Kannan 563b953224 DEV: Add 'backfill_etags_' to the method name since it also backfilling the etags 2019-02-19 21:54:35 +05:30
Vinoth Kannan 0472bd4adc FIX: Remove 'backfill_etags' keyword argument from 'uploads:missing' rake task
And etags backfilling code is optimized
2019-02-15 00:34:35 +05:30
Vinoth Kannan b5fbd7385f FIX: run the rake task only for uploads created before a day from inventory date 2019-02-14 17:53:08 +05:30
Vinoth Kannan a9a8855739 DEV: Get only matching records to backfill etags 2019-02-14 06:27:18 +05:30
Vinoth Kannan e2f7db5549 Fix typo 2019-02-14 05:56:30 +05:30
Vinoth Kannan 7b5931013a Update rake task to backfill etags from s3 inventory 2019-02-14 05:18:06 +05:30