1486707 - Firefox OWA inactivity timeout

Status: RESOLVED INCOMPLETE

(Core :: Networking: Cookies, defect, P4)

Description

[Dimas]

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:62.0) Gecko/20100101 Firefox/62.0
Build ID: 20180823143155

Steps to reproduce:

In our company we've updated Firefox from 43 to 52. Now many users complains about his session in the "Outlook web app" expires very soon, like 5-10 minutes, and before the update it was never less than 15 minutes.



Expected results:

I compared Firefox 43, Firefox 52, Firefox 60 and Internet Explorer 11. With IE11 and Ff 43 the session lasts much longer than Ff 52 and 60.

Comment 1

[Daniel Bodea [:danibodea]]

I have tried to leave a logged session of Outlook on "https://outlook.live.com/mail/inbox" page open and wait until something happens (like an auto log out), but I could not reproduce anything similar after waiting for more than 15 minutes.
I have tried this in multiple versions of the Firefox browser: Nightly v63.0a1, Release v62.0, Release v52.0.2, Release 52ESR, Release v60.1.0ESR, Release v43; neither of them reproduced losing the session in Outlook web app after 15 minutes f inactivity.

@Dimas, what exactly do you mean by "his session in the <Outlook web app> expires very soon"? Does the user get logged out?

Comment 2

[Dimas]

(In reply to Bodea Daniel [:danibodea] from comment #1)

> @Dimas, what exactly do you mean by "his session in the <Outlook web app>
> expires very soon"? Does the user get logged out?

Exactly, when users click the OWA tab after some minutes (<15) the login page appears, so they have to enter credentials again to see new mails.

But we don't use outlook.live.com, we use OWA for internal mail/exchange. After posting this issue to forums I see that other people reports the same problems with Firefox, but I don't know what OWA version they use.

In the about section of our OWA I can see:
- Version: 14.3.388.0
- Client access server version: 14.3.123.0
- Version of Microsoft Exchange server: 14.3.123.0

I know that it can be an OWA issue, but I wonder if some developer know about Firefox changes after 43 version that can affect.

Comment 3

[Daniel Bodea [:danibodea]]

I cannot investigate any further since we don't have OWA in our team.
I will set the component as (Firefox) Session Restore and let developers pick it up.
Please contact me if further testing is needed. 

Thank you for your contribution, Dimas!

Comment 4

[Honza Bambas (:mayhemer)]

Could this be another fallout of bug 368964?

Comment 5

[(no longer active)]

Yes, it could be, but it is hard to know without being able to reproduce.

Dimas, is there any chance that you can please help us narrow down when this problem started?  We have a tool called mozregression <https://mozilla.github.io/mozregression/> which allows you to find the exact code change that introduced the bug fairly quickly.  It usually takes 10-15 minutes to track a bug down, but in this case since you need to leave a session for a few minutes to try to see whether the bug appears, it would take more time.

That page contains a video guide on how to use the tool.  If you can help us fine what changed between Firefox 43 to 52, that would really help us towards fixing the problem.  Thanks in advance!  :-)

Comment 6

[Dimas]

I've been trying to limit the regression for two days and it's impossible given its random nature, I've no time to test so many builds. To mark a build as correct or incorrect I must do several tests and I can never be sure if it really works or next time it would fail.

I imagine that knowing the build that causes the error would be helpful, but in the next days I won't have time to do any more tests. I was hoping that someone might have clues as what might be causing the error, or maybe you already had conversations with Microsoft on the subject.

For my part, I will try to do tests again when I have time at work.

Comment 7

[(no longer active)]

Thank you Dimas for trying to help so far.  I will try reaching out to Microsoft to see if they have heard about any similar issues from their customers, or if they can confirm this bug on their end.

Comment 8

[Selena Deckelmann :selenamarie :selena]

(In reply to :Ehsan Akhgari from comment #7)
> Thank you Dimas for trying to help so far.  I will try reaching out to
> Microsoft to see if they have heard about any similar issues from their
> customers, or if they can confirm this bug on their end.

Any updates, Ehsan?

Comment 9

[(no longer active)]

None yet, I just pinged the thread again.

Comment 10

[Honza Bambas (:mayhemer)]

(Let's give this bug some more time, but in another 2 months I think it would be good to close as INCOMPLETE)

Comment 11

[Andrea Marchesini [:baku]]

See comment 10.

Comment 12

[Dimas]

Clearing needinfo