[SERVER-28876] fsyncLock results in corrupt disk snapshot Created: 20/Apr/17  Updated: 31/May/17  Resolved: 01/May/17

Status: Closed
Project: Core Server
Component/s: None
Affects Version/s: 3.2.12
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Dharshan Rangegowda Assignee: Mark Agarunov
Resolution: Duplicate Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Attachments: Zip Archive dataaftersnaprestart.zip     Zip Archive initialdata.zip     Zip Archive snapdata.zip    
Issue Links:
Duplicate
duplicates JAVA-2501 fsyncunlock does not work on secondar... Closed
Operating System: ALL
Participants:

 Description   

I am using server 3.2.12. FsyncLock() commands appear to corrupt the disk data such that a snapshot taken at this time does not work. Once the mongod process starts again on the snapshot disk it appears to ignore all the existing data files - none of the previous databases are shown.

I am able to reproduce the issue consistently. Here are my steps
1. Start with a sample data
2. db.fsyncLock()
3. Snapshot the disk
4. db.fsyncUnLock()

I have attached a .zip of mongodb data folder from my tests. Here is the login credential for initial data
mongo -u admin -p "K2NU3rNn0YBHX3p1" 127.0.0.1/admin

1. initialdata.zip - initial data
2. snapdata.zip - This is the data from the snapshot
3. dataaftersnaprestart - This is the data after restarting mongod on the snapshot



 Comments   
Comment by Mark Agarunov [ 01/May/17 ]

Hello dharshanr@scalegrid.net,

Thank you for the report. Looking over the output you've provided, this behavior appears to have the same cause as JAVA-2501. I have closed this ticket as a duplicate, so please follow JAVA-2501 for any updates on this issue.

Thanks,
Mark

Comment by Dharshan Rangegowda [ 20/Apr/17 ]

Also I have tried the same steps on 3.2.3 and it works fine.

Generated at Thu Feb 08 04:19:18 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.