[SERVER-29312] WiredTiger.wt: encountered an illegal file format or internal value Created: 22/May/17  Updated: 27/Jul/18  Resolved: 23/May/17

Status: Closed
Project: Core Server
Component/s: WiredTiger
Affects Version/s: 3.4.1
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Sophie Perrineau Assignee: Mark Agarunov
Resolution: Done Votes: 0
Labels: envns, rpo, rps, trcf, wtc
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Attachments: File SERVER-29312-repair-2.tar.gz     File WiredTiger.turtle     File WiredTiger.wt     Text File mongodb.log     File repair-SERVER-29312.tar.gz     Text File repair.log     Text File repair_secondserver.log     Zip Archive wd06.zip    
Operating System: Linux
Participants:

 Description   

Hello,
After a bad server/mongod shutdown, I am unable to restart Mongod. I'm 99% confident that my issue is that same as #SERVER-26103.
Repair command is not possible, because of the following error :
2017-05-22T13:26:25.739+0200 I CONTROL [initandlisten] options: { repair: true, storage:

{ dbPath: "/home/mongodb/" }

}
2017-05-22T13:26:25.758+0200 I - [initandlisten] Detected data files in /home/mongodb/ created by the 'wiredTiger' storage engine, so setting the active storage engine to 'wiredTiger'.
2017-05-22T13:26:25.762+0200 I STORAGE [initandlisten] exception in initAndListen: 72 Requested option conflicts with current storage engine option for directoryForIndexes; you requested false but the current server storage is already set to true and cannot be changed, terminating

I join to this ticket the wirertiger.wt files, and i hope that someone can repair it
Thanks for your help



 Comments   
Comment by Mark Agarunov [ 23/May/17 ]

Hello infrastructure,

Thank you for providing the additional information. Unfortunately, this error indicates that there was corruption on the disk. In this situation, my best recommendation would be to resync the affected node or restore from a backup if possible.

Please note that SERVER project is for reporting bugs or feature suggestions for the MongoDB server. For MongoDB-related support see our Technical Support page for additional resources.

Thanks,
Mark

Comment by Sophie Perrineau [ 23/May/17 ]

Hello,
I've extracted the files and run the rapir cmd on the second database.
This times, repair is failing :

  • It complains ajout sizeStorer.wt
  • then it salvages collections of the admin database
  • and then it fails to restore the first collection of a business database
    I join the complete log : repair_secondserver.log

Is there a way to force the content of this collection to empty in order to repair the other files ?

Comment by Mark Agarunov [ 23/May/17 ]

Hello infrastructure,

I'm glad to hear the repaired files worked. I've attached a repair attempt of the second set of files you've provided. Would you please extract these files and replace them in your $dbpath as with the previous set and let us know if it resolves the issue?

Thanks,
Mark

Comment by Sophie Perrineau [ 23/May/17 ]

The first DB is started without errors now. But we're still waiting for a solution for the other shard which has exactly the same problem (the damaged files are in the zip named wd06.zip).

Comment by Sophie Perrineau [ 23/May/17 ]

The first db seems to be OK (repair process ongoing). Could you please do the same operation for the second DB : the files are in the zip named wd06.zip

Comment by Mark Agarunov [ 22/May/17 ]

Hello infrastructure,

Thank you for the report. I've attached a repair attempt of the files you've provided. Would you please extract these files and replace them in your $dbpath and let us know if it resolves the issue?

Thanks,
Mark

Comment by Sophie Perrineau [ 22/May/17 ]

I have the same problem on a second server. Here are the corrupted files.
We also tried the reapir command with "--wiredTigerDirectoryForIndexes" option, it failed.

Generated at Thu Feb 08 04:20:28 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.