[jitsi-dev] Shutdown-Timeout


#1

Hey

Was there a specific reason to reduce the shutdown timeout from 15s to 3s?

I don't think that this was a good idea - we now force-quit sometimes while
the properties-file is being rewritten, leaving .part files. It would
probably be best to make the timeout dependent on the number of installed
accounts instead of a simple hardcoded value.

Ingo


#2

Hey

Was there a specific reason to reduce the shutdown timeout from 15s to 3s?

It was just too much ...

I don't think that this was a good idea - we now force-quit sometimes while
the properties-file is being rewritten, leaving .part files. It would
probably be best to make the timeout dependent on the number of installed
accounts instead of a simple hardcoded value.

I am not sure accounts have the biggest impact on file size (contact
caps seem to be the main culprit in my 10MB file) but I suppose we could
base that on the file size itself. Or we could just explicitly wait on
the write to finish (and have a separate timer for that) ... or
something along those lines.

Emil

···

On 22.01.13, 09:26, Ingo Bauersachs wrote:

--
https://jitsi.org


#3

While at that, why is it that Jitsi on OS X always quits with the .lock file remaining afterwards? Isn't that an indication that the shutdown procedure has not completed successfully?

···

On 22.01.2013, at 10:26, Ingo Bauersachs <ingo@jitsi.org> wrote:

Was there a specific reason to reduce the shutdown timeout from 15s to 3s?


#4

Was there a specific reason to reduce the shutdown timeout from 15s to

3s?

While at that, why is it that Jitsi on OS X always quits with the .lock

file

remaining afterwards? Isn't that an indication that the shutdown procedure
has not completed successfully?

I guess so. AFAIK the force-quit leaves a log message, do you see that?

Ingo

···

On 22.01.2013, at 10:26, Ingo Bauersachs <ingo@jitsi.org> wrote:


#5

Yes it is. But it's not supposed to be much of a problem. The file
contains an ip:port couple that we ping. If it doesn't respond fairly
quickly we continue.

···

On 22.01.13, 09:29, Lyubomir Marinov wrote:

On 22.01.2013, at 10:26, Ingo Bauersachs <ingo@jitsi.org> wrote:

Was there a specific reason to reduce the shutdown timeout from 15s
to 3s?

While at that, why is it that Jitsi on OS X always quits with the
.lock file remaining afterwards? Isn't that an indication that the
shutdown procedure has not completed successfully?

--
https://jitsi.org


#6

Hi,
I'm not sure if this is related but I have lost my contact lists a number of times during testing. I presume this is due to my frequent restarts of the client often facilitated by a process kill. This is still current as my sip contacts all disappeared over the last 24 hours.... again, which is a little frustrating :).

-James

···

-----Original Message-----

From: Emil Ivov [mailto:emcho@jitsi.org]

Sent: 22 January 2013 08:34
To: dev@jitsi.java.net
Cc: Lyubomir Marinov
Subject: [jitsi-dev] Re: Shutdown-Timeout

On 22.01.13, 09:29, Lyubomir Marinov wrote:

On 22.01.2013, at 10:26, Ingo Bauersachs <ingo@jitsi.org> wrote:

Was there a specific reason to reduce the shutdown timeout from 15s
to 3s?

While at that, why is it that Jitsi on OS X always quits with the
.lock file remaining afterwards? Isn't that an indication that the
shutdown procedure has not completed successfully?

Yes it is. But it's not supposed to be much of a problem. The file contains an ip:port couple that we ping. If it doesn't respond fairly quickly we continue.

--
https://jitsi.org


#7

Hey James,

Hi, I'm not sure if this is related but I have lost my contact lists
a number of times during testing. I presume this is due to my
frequent restarts of the client often facilitated by a process kill.
This is still current as my sip contacts all disappeared over the
last 24 hours.... again, which is a little frustrating :).

It could be related although, unfortunately the problems with the
contact list date from quite a while before that.

We are hoping that one deay we'll have time enough to move to an sqlite
or an hsql backend for this.

Cheers,
Emil

···

On 22.01.13, 22:00, James M. Martin wrote:

-James

-----Original Message----- From: Emil Ivov [mailto:emcho@jitsi.org]
Sent: 22 January 2013 08:34 To: dev@jitsi.java.net Cc: Lyubomir
Marinov Subject: [jitsi-dev] Re: Shutdown-Timeout

On 22.01.13, 09:29, Lyubomir Marinov wrote:

On 22.01.2013, at 10:26, Ingo Bauersachs <ingo@jitsi.org> wrote:

Was there a specific reason to reduce the shutdown timeout from
15s to 3s?

While at that, why is it that Jitsi on OS X always quits with the
.lock file remaining afterwards? Isn't that an indication that the
shutdown procedure has not completed successfully?

Yes it is. But it's not supposed to be much of a problem. The file
contains an ip:port couple that we ping. If it doesn't respond fairly
quickly we continue.

-- https://jitsi.org

--
https://jitsi.org