[GH-ISSUE #239] Re-attach to an overlay after the sandbox is closed #171

Closed
opened 2026-05-05 05:13:43 -06:00 by gitea-mirror · 5 comments
Owner

Originally created by @netblue30 on GitHub (Jan 20, 2016).
Original GitHub issue: https://github.com/netblue30/firejail/issues/239

From wordpress:

Hi,

Is it possible to re-attach to an overlay after the sandbox is closed? Now, a new overlay is always created, but I would like to continue where I left of.

Thanks,
eli

Originally created by @netblue30 on GitHub (Jan 20, 2016). Original GitHub issue: https://github.com/netblue30/firejail/issues/239 From wordpress: Hi, Is it possible to re-attach to an overlay after the sandbox is closed? Now, a new overlay is always created, but I would like to continue where I left of. Thanks, eli
gitea-mirror 2026-05-05 05:13:43 -06:00
Author
Owner

@dshmgh commented on GitHub (Feb 13, 2016):

I have experimented with the overlayfs mount and if I manually remount over a previous sandbox, the merge appears to work properly. If an option is added --reoverlay and then that creates the /home/$USER/.firejail/persist dir and that dir is used instead of /home/$USER/.firejail/$PID for the overlay then I think this enhancement would work. But does this reuse of a previous sandbox somewhat defeat the purpose of using firejail?

I'll get some code for this change to you at some point.

<!-- gh-comment-id:183718968 --> @dshmgh commented on GitHub (Feb 13, 2016): I have experimented with the overlayfs mount and if I manually remount over a previous sandbox, the merge appears to work properly. If an option is added --reoverlay and then that creates the /home/$USER/.firejail/persist dir and that dir is used instead of /home/$USER/.firejail/$PID for the overlay then I think this enhancement would work. But does this reuse of a previous sandbox somewhat defeat the purpose of using firejail? I'll get some code for this change to you at some point.
Author
Owner

@netblue30 commented on GitHub (Feb 13, 2016):

I'm really interested in this feature, so if you send me some code I'll merge it in. The reason I've been sleeping on it is remounting the overlay is a huge security problem. For example, in the overlay you modify /etc/shadow, then start firejail and become root. I'll have to restrict --reoverlay to root user.

<!-- gh-comment-id:183771766 --> @netblue30 commented on GitHub (Feb 13, 2016): I'm really interested in this feature, so if you send me some code I'll merge it in. The reason I've been sleeping on it is remounting the overlay is a huge security problem. For example, in the overlay you modify /etc/shadow, then start firejail and become root. I'll have to restrict --reoverlay to root user.
Author
Owner

@Sidnioulz commented on GitHub (Feb 21, 2016):

@netblue30 would the issue exist if user namespaces were in place?

<!-- gh-comment-id:186715281 --> @Sidnioulz commented on GitHub (Feb 21, 2016): @netblue30 would the issue exist if user namespaces were in place?
Author
Owner

@netblue30 commented on GitHub (Feb 21, 2016):

Overlayfs and user namespace should be two different subsystems.

<!-- gh-comment-id:186820936 --> @netblue30 commented on GitHub (Feb 21, 2016): Overlayfs and user namespace should be two different subsystems.
Author
Owner

@netblue30 commented on GitHub (Sep 6, 2016):

Implemented in 0.9.42~rc2

<!-- gh-comment-id:244933084 --> @netblue30 commented on GitHub (Sep 6, 2016): Implemented in 0.9.42~rc2
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/firejail#171
No description provided.