Bug 512369 - handle detection of corrupted user.json
Summary: handle detection of corrupted user.json
Status: REOPENED
Alias: None
Product: Orion (Archived)
Classification: ECD
Component: Server (show other bugs)
Version: unspecified   Edit
Hardware: PC Mac OS X
: P4 normal (vote)
Target Milestone: ---   Edit
Assignee: Grant Gayed CLA
QA Contact:
URL:
Whiteboard:
Keywords: triaged
Depends on:
Blocks:
 
Reported: 2017-02-17 08:38 EST by Grant Gayed CLA
Modified: 2018-04-09 12:10 EDT (History)
1 user (show)

See Also:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Grant Gayed CLA 2017-02-17 08:38:28 EST 
Recent server work revealed that it's easy for a user.json file to become corrupted if the server gets killed improperly (eg.- a VM reboot).  Once this happens the user is permanently unable to log in.  In the recent cases the user.json files were present but had size 0, so presumably got interrupted during a write.

We handle the case of a missing user.json fine, by creating a new one.  But in the present-but-empty case we fail miserably.  At a minimum we should detect this present-but-empty case and handle it (just delete-then-recreate it?).  This approach may be advisable any time a corrupted user.json is detected.
Comment 1 Steve Northover CLA 2017-03-30 16:57:09 EDT 
This sounds like Java server work so lower priority for Orion 15.
Comment 2 Michael Rennie CLA 2018-04-09 09:36:52 EDT 
Closing as part of a mass clean up of inactive bugs. Please reopen if this problem still occurs or is relevant to you. 

For more details see: https://dev.eclipse.org/mhonarc/lists/orion-dev/msg04114.html
Comment 3 Grant Gayed CLA 2018-04-09 12:10:15 EDT 
Reopening and adding 'triaged' keyword.  Need to ensure the node server handles this reasonably.