My forum old images not see


(T. H. Wright) #21

I am also experiencing this transparent.png bug. I had updated from 2.1 to 2.2 today without realizing the issue and have run the following.

[2] pry(main)> OnceoffLog.where(job_name:"RecoverPostUploads").count
=> 1

I do not have Cloudflare enabled and pictures are stored locally, though the server is accessed by a proxy.

Before doing the above, I had run an rsync to add the missing images back to the uploads directory; this resolved some problems but not those with transparent.png being listed. The most recent post affected was from August 12.

Do I simply need to wait an hour as Sam had mentioned, or do I need to do anything else?

Uncertain

I wasn’t sure if I should start a new thread or piggyback off of this one; I opted to piggyback.


(Alan Tan) #22

Can you try running the following?

./launcher enter app
rails c
require_dependency "upload_recovery"
UploadRecovery.new.recover

(Hakan) #23

Ekran%20Al%C4%B1nt%C4%B1s%C4%B1


(T. H. Wright) #24
root@www:/var/www/discourse# rails c
[1] pry(main)> require_dependency "upload_recovery"
=> nil
[2] pry(main)> UploadRecovery.new.recover
=> nil
[3] pry(main)> exit

That seemed to do the job for all but one image which I pulled from a backup and reuploaded. Any ideas why large images would no longer be lightboxed?

I ask as I am utilizing the Tiles and Slick theme components and these are no longer functioning.


(Alan Tan) #25

Regarding the light box, please try and reproduce on try.discourse.org and open another topic if you’re able to do so.


(Sam Saffron) #26

This is odd @tgxworld, the job ran so why did he need to run it again?

I guess this just validates how important it is to integrate this into the migration system


(Alan Tan) #27

I’m not sure actually without being able to look at the records before the recovery job has been executed. The recovery job I added was meant to fix the bug I introduced in the 2.1 release but it looks like @thwright was having problems with posts created on Aug 12 which may be caused by the other bug.


(Alan Tan) #28

I think something is off with the onceoff system. Records in the OnceoffLog said it ran the job but running it manually fixed the broken records for one of the site I’m looking at.

[6] pry(main)> UploadRecovery.new(dry_run: true).recover(Post.where("baked_at >= ?", SiteSetting.purge_deleted_uploads_grace_period_days.days.ago))
# Displays two broken records
=> nil
[7] pry(main)> OnceoffLog.last
=> #<OnceoffLog:0x00007f599ebbc6b0 id: 34, job_name: "RecoverPostUploads", created_at: Wed, 19 Sep 2018 07:06:21 UTC +00:00, updated_at: Wed, 19 Sep 2018 07:06:21 UTC +00:00>
[8] pry(main)> Jobs::RecoverPostUploads.new.execute_onceoff({})
=> nil
[9] pry(main)> UploadRecovery.new(dry_run: true).recover(Post.where("baked_at >= ?", SiteSetting.purge_deleted_uploads_grace_period_days.days.ago))
=> nil
[10] pry(main)>