Various POST requests get stuck forever after bug 1772201
Categories
(Core :: Networking: HTTP, defect, P1)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox-esr91 | --- | unaffected |
| firefox-esr102 | --- | unaffected |
| firefox103 | --- | unaffected |
| firefox104 | --- | unaffected |
| firefox105 | --- | affected |
People
(Reporter: denschub, Unassigned)
References
(Regression)
Details
(Keywords: regression)
Attachments
(1 file)
|
5.37 MB,
application/x-gzip
|
Details |
Over the weekend, I've noticed a lot of applications randomly failing. I couldn't press a button in a KVM controller's Web UI, I couldn't log into my online banking, and a bunch of stuff just randomly failed or timeouted.
There's no indication in the devtools, the request just stays there without a status. I've attached a network log from a firefox instance with nothing but a single tab trying to do one of those requests. The request in this case was a POST to https://pikvm-desktop.kvm.schub.bar/api/atx/click?button=power. This is a LAN-only URL, but obviously I can't share a log of my online banking publicly... :)
This is a regression from bug 1772201. mozregression points me towards D153741, but since there are multiple commits in the same push, I'm not sure if that's accurate.
kershaw, I know that bug 1772201 has a lot of regressions filed against it already, but this seems like a different issue. Random sites/apps fail in random POST requests, reproducible 100% of the time for me, in a completely fresh profile with no settings touched whatsoever.
I'm setting this P1/S1. Since this isn't just breaking a LAN-device, but also the login into my online banking at Kontist.com, there is potential for a lot of fallout here. Also, in the case of Kontist, I didn't even see any indication in the UI - it just loaded forever. As a bonus fun fact, I couldn't even file this bug in Nightly, and had to switch to Beta. The POST just loaded forever.
|
Reporter | |
Comment 1•2 years ago
|
||
Setting correct tracking flags, keeping ? for 105 as this is a serious regression.
|
|
Reporter | |
Comment 2•2 years ago
|
||
Also, forgot to ni? kershaw since I had to switch browsers... :)
|
|
Reporter | |
Comment 3•2 years ago
|
||
Just realized there is already a patch in https://bugzilla.mozilla.org/show_bug.cgi?id=1783567#c5, so this might be on it's way to being resolved. I'll verify as soon as that patch is in central.
|
||
Comment 4•2 years ago
|
||
Based on the log, I can confirm this is the same as bug 1783567.
Thanks.
|
|
Reporter | |
Updated•2 years ago
|
Description
•