Page MenuHomeMiraheze

Handling of hkrail.miraheze.org after database recovering
Closed, ResolvedPublic

Description

NOTE: This is not a db141 wiki recreation request, the wiki has already been recreated, please don't merge it into T​10015.

Per your team's announcement of the plan of recovering the old db141 database and then importing edits from a backup of the recreated wikis, it is in my interest to instead have the recreated wiki backup as the base of my wiki and I manually extract part of the wiki dump from my old data to either manually copy and paste in (for recreated pages) or import dump some other parts of it.

Per conversation with @Agent_Isai, it will probably require deleting and recreating the wiki again to achieve this. As this will probably be stalled until the wikis are back up, I'm filing this to Phab so that it won't get completely lost.

I believe these would be the steps to do after wiki recovery:

  • Wiki recovery
    • XML Dump of Nov pre-crash version
    • XML Dump of Dec recreated version (sent via Discord DM fr Cosmic to Lucifer)
  • Backup and delete wiki
  • Recreate wiki (details provided again below FYI)
    • Relink images
  • Import Dec 18 backup for the recreated (3.0) wiki
  • (my part) Uploading data dump requests for pages I need to import again (or if I can import them myself I will do so)

Additionally I would ask if it is possible to recover image uploading history when relinking the images, coz it will now permanently say it is (first) uploaded by a bot script.

Thanks!


The following is information about the wiki in the same format as the wiki recreation request for your convenience:

Wiki subdomainhkrail.miraheze.org
Wiki usernameLuciferianThomas
Wiki name香港鐵路百科
Languagezh (Chinese)
VisibilityPublic
CategorizationUncategorized

Event Timeline

LuciferianThomas created this task.

Update: It appears there are exceptions site-wide.

Update: It appears there are exceptions site-wide.

Fixed. As Discussed on Discord I also ran the AssignImportedEdits script as well.

Universal_Omega updated the task description. (Show Details)

All parts that require action from us as indicated in the description is now done. Please reopen if there is something else. Thanks!

A ton of pages doesn't work after remapping broken page types to plain text, I see a ton of page with mismatched page types refusing to work, mostly Scribunto modules, TemplateStyles CSS, and as well the site MediaWiki:Common.css.

If needed or possible, should we try resetting it again but with the required extensions installed correctly first before importing pages? It seems to be way easier than fixing hundreds of pages with inconsistent page types targets to be fixed.

Further edit: Many pages are also imported to the wrong title, a reset is likely needed.

Updated task request after discussion with @Universal_Omega:-

A full cleanup for page content models is needed and all logs and revisions after specifically 2023-01-14T01:53:00Z to be pruned.