Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[1.X][2.X] Connection occasionally times out and does not recover → app becomes unable to load sites #2

Closed
mtigas opened this issue Apr 17, 2012 · 15 comments

Comments

@mtigas
Copy link
Member

mtigas commented Apr 17, 2012

Workaround: Force quit app (tap home to quit app, double-tap home, hold down on app icon in tray, and "delete" app) and then restart app.


Hard to recreate issue when running app for more than a few minutes.

Seems to happen regardless of if app has gone into background or if network has gone up/down. (In fact, app does send HUP to Tor in those circumstances, causing Tor to reload & build new connections.)
#3 would help (and is related), but this is a more specific issue than that.

@frou
Copy link

frou commented Apr 27, 2012

This happens very frequently for me.

The app is in the foreground, the Auto-Lock (2 mins) kicks in, then when I go back and unlock the iPad, odds are that functionality is unrecoverable and the app needs killed. The iPad is on a solid WiFi connection.

@mtigas
Copy link
Member Author

mtigas commented Apr 27, 2012

Based on @frou and some other folks I’ve heard from, looks like this is specific to (or at least, primarily related to) auto-lock.

Manually locking and then coming back has seemed to work fine in my tests (but probably affected by some internal auto-lock timeout, too), but I’ll have to look at this more.

@nhodges
Copy link

nhodges commented May 3, 2012

This definitely seems to be the case, I can confirm one, probably two iPhone 4S's losing the ability to browse after an autolock.

@kyliekarnage
Copy link

No need to wait for autolock. My 4S does it any time you leave the app.

mtigas added a commit that referenced this issue Sep 6, 2012
* checkTor fires every five seconds
* when checkTor fires, it sets a 1sec timeout
* this timeout basically means that the control port is hung up (so we attempt to restart tor)
@mtigas
Copy link
Member Author

mtigas commented Sep 8, 2012

Rewrote a lot of the code that handles the "glue" between the Tor bits and the OnionBrowser iOS bits: d470cff. Currently have a minor update waiting in the review queue (submitted 9/5, probably will be through on 9/12), but I'll submit the new (1.3.0) version after that and see if that helps some.

@mtigas mtigas mentioned this issue Jul 29, 2013
@jbierling
Copy link

Just bought the app and this is still is happening on my mini. Force-quitting seems to help.

@Bobsson
Copy link

Bobsson commented Sep 16, 2013

Ditto to @jbierling. Given the page load speed of the TOR network, it's convenient to be able to put the phone down while the page loads, then pick it up again a minute or two later. Unfortunately, this often results in needing to force-quit and start all over.

@John-Khart
Copy link

Hi. It seems that this problem is still happening. Tested and verified on iOS 7.1.1 on iPhone 5s, the app still has to be forced closed and restarted should network reconnections (eg. switching from 3G to 4G) happen. No issues with phone locking though.

mtigas added a commit that referenced this issue Nov 12, 2014
…nt has stopped working

well, i'm pretty sure this catches tor death. (ref #2)
@MarkCallow
Copy link

Still happens on iOS 9. It is the biggest problem I have with OB. The friendlier error just makes one want to scream: "if they know about this, why don't they fix it?"

@mtigas mtigas changed the title Sometimes loses connection → becomes unable to load sites [1.X][2.X] Connection occasionally times out and does not recover → app becomes unable to load sites Apr 20, 2017
@mtigas mtigas added this to Tor controller in pre-2.0 development Apr 20, 2017
@DavidMOliver
Copy link

I want to add this to milestone 2.1 for tracking throughout milestones 2.1 - 2.9

@tladesignz
Copy link
Contributor

@DavidMOliver: As discussed frequently on telcos: We're completely dependent on Tor Project with this. May need some more personal sessions with the Tor folks. (Did somebody say travel budget? ^^) So to put this in a milestone doesn't make too much sense at least in terms of the GitHub issue tracker. Multiple milestones are not possible anyway, here.

The propsal is another story, though.

@DavidMOliver
Copy link

@tladesignz no question we need help. We'll work this out in a way that is not "our development work" and make the dependency clear in the proposal

@mtigas
Copy link
Member Author

mtigas commented Oct 15, 2018

Quick update on this: A lot of work has been done by the core tor folks to try to at least allow us to kill tor (separately from killing the app process) and start it back up later. See https://trac.torproject.org/projects/tor/ticket/25510 (this is basically the "in-process restart" stuff mentioned in sub-tickets linked from there).

Tor has been updated to something newer than 0.3.3.X, which means we get this API.

I've enabled some experimental changes that @tladesignz put in, and left them very rough. 13fe387 It's super rough: the "Connecting..." dialog doesn't appear again when the app comes back to foreground so there's a good chance that the app will still have a silent network hang if something bad happens. There's also good chance that this is super unstable and crashes. But it's worth putting it out in the TestFlight builds, IMO.

@mtigas mtigas added this to To do in Onion Browser 2019 Jan 14, 2019
@tladesignz tladesignz added this to the 2.1.0 milestone Jan 24, 2019
@tladesignz tladesignz self-assigned this Jan 29, 2019
@tladesignz
Copy link
Contributor

Finally with release 2.2.0 there's going to be Tor 0.3.5.8 contained which will should fix this for good.

However, I'll leave this issue open and forward it to the next release 2.3.0 for collection of feedback.

Additionally, there's one small issue left, as we currently don't get progress/connection feedback on restarts from the Tor binary, which could lead to situations, when the background restart doesn't work as expected.

So, not completely resolved, but mostly.

@tladesignz tladesignz modified the milestones: 2.2.0, 2.3.0 Mar 5, 2019
@tladesignz tladesignz removed this from the 2.2.1 (was 2.3.0) milestone Jun 20, 2019
@tladesignz tladesignz added this to the 2.3.0 (was 2.4.0) milestone Jun 20, 2019
@tladesignz
Copy link
Contributor

I'll close this now, as we didn't hear any complaints lately. The restart of Tor seems to be quite stable now.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
Onion Browser 2019
  
Milestone A1
pre-2.0 development
Tor controller
Development

No branches or pull requests

10 participants