Closed Bug 1768250 Opened 2 years ago Closed 2 years ago

Using www.youtube.com thumbnails stop being retrieved and youtube goes "offline"

Categories

(Core :: Networking, defect, P2)

Firefox 100
defect

Tracking

()

RESOLVED FIXED
107 Branch
Tracking Status
firefox107 --- fixed

People

(Reporter: gregdyck, Assigned: kershaw, NeedInfo)

References

Details

(Whiteboard: [necko-triaged])

Attachments

(2 files)

Steps to reproduce:

Go to www.youtube.com and start browsing videos. This started to occur after Firefox 100.0 (64 bit) was installed. At 99.0.1 (64 bit) all worked correctly.

Actual results:

Eventually the preview for some of the videos stays grey. After a longer time a screen saying "Your offline. Check your internet connection." is displayed.

Expected results:

The previews should have been resolved and displayed. There is an internet connection.

The Bugbug bot thinks this bug should belong to the 'Core::Networking' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Component: Untriaged → Networking
Product: Firefox → Core

Hi,

Could you try to capture a http log?
If this is easy to reproduce, could you also try to use mozregression to help us find out the root cause?

Thanks.

Flags: needinfo?(gregdyck)
Attached file log.txt-main.15152.moz_log.zip (deleted) —

I have attached a log that I started after the problem began to occur. The problem occurs regularly, but there appears to be more involved to trigger it than I thought. I will keep trying to recreate with logging active from the very start.

Flags: needinfo?(gregdyck)

Thanks for the log.
This looks like related to 0RTT. Could you try to go to about:config and set security.tls.enable_0rtt_data to false and try again?
It'd be great if you can also provide more details about STR.

Flags: needinfo?(gregdyck)

I have been unable to get the failure to occur if I start http logging up front, but have had it reoccur when it is not started. Perhaps logging is changing timing, and/or causing serialization to occur which prevents the failure from starting?

I have disabled config security.tls.enable_0rtt_data as requested, and will see if the failure has been resolved.

I don't know what you want when you asked "can also provide more details about STR."

Flags: needinfo?(gregdyck)

FYI, the problem has not reoccurred since I changed the setting of security.tls.enable_0rtt_data to disabled.

The severity field is not set for this bug.
:kershaw, could you have a look please?

For more information, please visit auto_nag documentation.

Flags: needinfo?(kershaw)
Flags: needinfo?(kershaw)
Whiteboard: [necko-priority-review]

The severity field is not set for this bug.
:dragana, could you have a look please?

For more information, please visit auto_nag documentation.

Flags: needinfo?(dd.mozilla)
Severity: -- → S3
Priority: -- → P2
Whiteboard: [necko-priority-review] → [necko-priority-queue]

gregdyck, are you using any firewall? For example, Sophos has a known issue with 0RTT.

Dennis, do you know of any change related to 0RTT in Fx 100? In bug 1770742 I have fixed a problem with our retry mechanism. n the log from this bug I do not see SSL_ERROR_BAD_MAC_ALERT error only NET_RESET.

Flags: needinfo?(gregdyck)
Flags: needinfo?(djackson)
Flags: needinfo?(dd.mozilla)

There were no changes to 0RTT in NSS 3.77 which shipped with Fx 100.

Flags: needinfo?(djackson)

I can confirm the problem is gone for me after disabling security.tls.enable_0rtt_data
Also, I have Sophos installed

Yes, Sophos was pre-installed on my system by corporate policy. It can not be removed.

This is the mitigation when a firewall disallows Firefox to use 0RTT. A new pref is added to control whether to always use 0RTT for HTTP/2.
When the pref is false, nsHttpTransaction::Do0RTT is called before trying 0RTT.

Assignee: nobody → kershaw
Status: UNCONFIRMED → ASSIGNED
Ever confirmed: true
Whiteboard: [necko-priority-queue] → [necko-priority-queue][necko-triaged]

Lower the severity a bit, since it seems the failure of 0RTT is not caused by a firewall or a middle box.
We are also discussing a better retry algorithm to fix this kind of issues once for all.

Severity: S3 → S4
Pushed by kjang@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/d6caea480d4d Don't always do 0RTT for HTTP/2, r=necko-reviewers,dragana
Flags: needinfo?(kershaw)
Whiteboard: [necko-priority-queue][necko-triaged] → [necko-triaged]
Pushed by kjang@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/48faf80010b4 Don't always do 0RTT for HTTP/2, r=necko-reviewers,dragana
Status: ASSIGNED → RESOLVED
Closed: 2 years ago
Resolution: --- → FIXED
Target Milestone: --- → 107 Branch
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: