Red Hat Bugzilla – Full Text Bug Listing
|Summary:||Propose that you turn on PrivateTmp=true in service file for Zarafa|
|Product:||[Fedora] Fedora||Reporter:||Daniel Walsh <dwalsh>|
|Component:||zarafa||Assignee:||Robert Scheck <redhat-bugzilla>|
|Status:||CLOSED CANTFIX||QA Contact:||Fedora Extras Quality Assurance <extras-qa>|
|Fixed In Version:||Doc Type:||Bug Fix|
|Doc Text:||Story Points:||---|
|Last Closed:||2012-02-06 15:54:49 EST||Type:||---|
|oVirt Team:||---||RHEL 7.3 requirements from Atomic Host:|
|Bug Depends On:|
Description Daniel Walsh 2012-01-17 10:10:03 EST
I would like to propose using PrivateTmp for Zarafa, to make it more secure and avoid users from being able to potentially effect it. http://fedoraproject.org/wiki/Features/ServicesPrivateTmp
Comment 1 Robert Scheck 2012-01-17 10:14:12 EST
I am tempted to use /var/lib/zarafa/autoresponder (or similar) instead of /tmp, because /tmp is on recent Fedora anyway on a ramfs, thus /tmp gets emptied out during reboots and all set up autoresponder history data are gnone, too. It's anyway not really temporary data, but application created data that should exist as long as it's needed. Objections?
Comment 2 Daniel Walsh 2012-01-17 11:28:52 EST
No that would be perfect. I don't believe any privileged process should use /tmp. /tmp is for users /var/lib and /run are for system processes...
Comment 3 Daniel Walsh 2012-02-06 15:45:52 EST
Any change on this bug. We are coming up to Feature Freeze, and would like some comment on this bug. If you do not believe this application uses /tmp than please comment on this and close the bug. If you believe this application needs to use /tmp to communicate with other applications or users then you can close this bug with that comment. If your app does not use systemd, then close this bug with that comment. If you have no idea, then please add a comment, and change the bug to assigned. I need to update the status on this feature. Thanks for your help.
Comment 4 Robert Scheck 2012-02-06 15:54:49 EST
Zarafa does not yet use systemd. The usage of /tmp is not related to the initscript, however there might be a /tmp usage of the initscript which I need to discuss with upstream anyway. But can't be solved without systemd switching for now anyway.