I have a production environment which is running freeradiusd 2.1.8 and last night in the logs I see the following message Sat Jan 1 20:11:24 2011 : Error: Mon Jan 10 17:04:58 2011 : Info: Exiting normally. No one was on the box doing anything... I was looking into this issue with google and came across a thread back in Nov 2009 about an issue a user was experiencing with radiusd 2.1.8, and this user send some gdb dumps to the development team... I can't seem to recreate the issue as quickly as he does (plus my server is in production) but I didn't see any follow up if this is a known bug? Is this fixed in 2.1.10? Thanks, Brian
Brian Carpio <bcarpio@broadhop.com> writes:
I have a production environment which is running freeradiusd 2.1.8 and last night in the logs I see the following message
Sat Jan 1 20:11:24 2011 : Error: Mon Jan 10 17:04:58 2011 : Info: Exiting normally.
No one was on the box doing anything... I was looking into this issue with google and came across a thread back in Nov 2009 about an issue a user was experiencing with radiusd 2.1.8,
2.1.8 was released Dec 30 2009. But freeradius version numbers are usually updated early in the developement cycle. The code discussed in Nov 2009 was a development version between 2.1.7 and 2.1.8. IIRC the issue discussed was fixed well before the 2.1.8 release. 2.1.8 was very stable.
and this user send some gdb dumps to the development team... I can't seem to recreate the issue as quickly as he does (plus my server is in production) but I didn't see any follow up if this is a known bug? Is this fixed in 2.1.10?
There are lots of fixes in 2.1.10, which has been as stable as 2.1.8 was for us (2.1.9 was not...). I would have upgraded before investing any more time debugging the problem. Maybe even to current git, as 2.1.11 is really close according to Alan DeKok. It may not fix your problem, but at least it will make your debugging useful to the developers. 2.1.8 is getting sort of outdated. Bjørn
Brian Carpio wrote:
No one was on the box doing anything… I was looking into this issue with google and came across a thread back in Nov 2009 about an issue a user was experiencing with radiusd 2.1.8, and this user send some gdb dumps to the development team… I can’t seem to recreate the issue as quickly as he does (plus my server is in production) but I didn’t see any follow up if this is a known bug? Is this fixed in 2.1.10?
It's something which is seen only on FreeBSD. As far as I can tell, it's a race condition in the threading code. It's difficult to track down and debug, unfortunately. Given the level of complexity of the current code, it's probably best to re-write it, to be simpler and cleaner. That way the bugs will be easier to track down. Alan DeKok.
Alan DeKok-2 wrote:
It's something which is seen only on FreeBSD. As far as I can tell, it's a race condition in the threading code. It's difficult to track down and debug, unfortunately.
Given the level of complexity of the current code, it's probably best to re-write it, to be simpler and cleaner. That way the bugs will be easier to track down.
Does this mean we shouldn't expect this bug to be fixed in new releases, unless the code gets rewritten? / Danial -- View this message in context: http://freeradius.1045715.n5.nabble.com/Freeradiusd-2-1-8-tp3346801p3415793.... Sent from the FreeRadius - User mailing list archive at Nabble.com.
FBSD allows one to easily change the scheduler, MAYBE that would help? -----Original Message----- From: freeradius-users-bounces+ggatten=waddell.com@lists.freeradius.org [mailto:freeradius-users-bounces+ggatten=waddell.com@lists.freeradius.org] On Behalf Of Danial Sent: Wednesday, March 09, 2011 9:48 AM To: freeradius-users@lists.freeradius.org Subject: Re: Freeradiusd 2.1.8 Alan DeKok-2 wrote:
It's something which is seen only on FreeBSD. As far as I can tell, it's a race condition in the threading code. It's difficult to track down and debug, unfortunately.
Given the level of complexity of the current code, it's probably best to re-write it, to be simpler and cleaner. That way the bugs will be easier to track down.
Does this mean we shouldn't expect this bug to be fixed in new releases, unless the code gets rewritten? / Danial -- View this message in context: http://freeradius.1045715.n5.nabble.com/Freeradiusd-2-1-8-tp3346801p3415793.... Sent from the FreeRadius - User mailing list archive at Nabble.com. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html <font size="1"> <div style='border:none;border-bottom:double windowtext 2.25pt;padding:0in 0in 1.0pt 0in'> </div> "This email is intended to be reviewed by only the intended recipient and may contain information that is privileged and/or confidential. If you are not the intended recipient, you are hereby notified that any review, use, dissemination, disclosure or copying of this email and its attachments, if any, is strictly prohibited. If you have received this email in error, please immediately notify the sender by return email and delete this email from your system." </font>
Gary Gatten wrote:
FBSD allows one to easily change the scheduler, MAYBE that would help?
I've tried changing scheduler, but alas to no avail. It was worth a shot though. / Danial -- View this message in context: http://freeradius.1045715.n5.nabble.com/Freeradiusd-2-1-8-tp3346801p4231968.... Sent from the FreeRadius - User mailing list archive at Nabble.com.
Danial wrote:
Does this mean we shouldn't expect this bug to be fixed in new releases, unless the code gets rewritten?
Or unless someone tracks it down and submits a patch. It's open source software. There's no need to wait for the bug to magically get fixed. People *can* contribute. Alan DeKok.
participants (5)
-
Alan DeKok -
Bjørn Mork -
Brian Carpio -
Danial -
Gary Gatten