Freeradius 1.1.5 on RH 7.3 crashes
Dear All, I successfully used freeradius-1.0.5 on our RedHat 7.3 system for a couple years to authenticate wireless and VPN clients. I decided to upgrade to the version 1.1.5 to enable the wireless connection on a few Windows Vista clients. I downloaded the source freeradius-1.1.5.tar.gz and customized the provided freeradius.spec. Essentially in the configure I added: --with-openssl-includes=/usr/local/ssl/include \ --with-openssl-libraries=/usr/local/ssl/lib \ (as openssl-0.9.7i is installed in /usr/local/ssl/lib) and removed: --with-system-libtool \) and succefully produced the rpm for my system. I installed it and it worked even for 6, 7 hours (enabling the wireless clients to work, included the Vista clients) but then invariably it crashed. Apparently, (from the various logs attached) it always crashed after the configuration reload, but it is not evident if immedialtely after the reload or some time later. The crash can some time be initiated by sending kill -HUP at the radiusd process. Trying to minimze the frequency of the crashed I set: max_request_time = 10 but without any change in the behaviour. Any suggestion on how to solve the problem would be appreciated. Best regards Dario P.S.: Attached is a partial log (otherwise was too big), only username and password are masked). I can provide longer logs on request. -- ______________________________________________________________ Dario Palmisano ICGEB Computer System & Network Administrator Tel: +39 040 3757330 Fax: +39 040 226555 E-Mail: Dario.Palmisano@icgeb.org International Centre for Genetic Engineering and Biotechnology Area Science Park, Padriciano 99, I-34012 Trieste, ITALY ______________________________________________________________
Hi,
clients to work, included the Vista clients) but then invariably it crashed. Apparently, (from the various logs attached) it always crashed after the configuration reload, but it is not evident if immedialtely after the reload or some time later.
are you 'HUP'ing the server at those times to reload the config? I too have noticed change in behaviour with 1.1.5 regarding HUP of the server. 1.1.4 was fine, but a HUP on 1.1.5 can kill the process alan
Hi Alan, no, many crashed were spontaneus (I do not know the internals of radiusd, but from time to time It reloads, maybe as response to a passwd file changed or so). Then I tried to send an HUP signal, after the server authenticated some clients, and few seconds (20-40) later the server crashed. Unfortunately the log does not show the activity during the crash, maybe someone can suggest me how to get more detailed log. Thanks Dario On Wednesday 14 March 2007 16:17, A.L.M.Buxey@lboro.ac.uk wrote:
Hi,
clients to work, included the Vista clients) but then invariably it crashed. Apparently, (from the various logs attached) it always crashed after the configuration reload, but it is not evident if immedialtely after the reload or some time later.
are you 'HUP'ing the server at those times to reload the config? I too have noticed change in behaviour with 1.1.5 regarding HUP of the server. 1.1.4 was fine, but a HUP on 1.1.5 can kill the process
alan - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
-- ______________________________________________________________ Dario Palmisano ICGEB Computer System & Network Administrator Tel: +39 040 3757330 Fax: +39 040 226555 E-Mail: Dario.Palmisano@icgeb.org International Centre for Genetic Engineering and Biotechnology Area Science Park, Padriciano 99, I-34012 Trieste, ITALY ______________________________________________________________
Dario Palmisano wrote:
no, many crashed were spontaneus (I do not know the internals of radiusd, but from time to time It reloads, maybe as response to a passwd file changed or so). Then I tried to send an HUP signal, after the server authenticated some clients, and few seconds (20-40) later the server crashed.
There is a bug filed on bugs.freeradius.org. Sending the server HUP signals just isn't safe until it's fixed. It may be possible to fix it in 1.1.6...
Unfortunately the log does not show the activity during the crash, maybe someone can suggest me how to get more detailed log.
The issue isn't the activity during the crash, but the activity during the HUP. You have a module that's taking a LONG time to process a request, and it's preventing the HUP handling from proceeding correctly. Alan DeKok. -- http://deployingradius.com - The web site of the book http://deployingradius.com/blog/ - The blog
A.L.M.Buxey@lboro.ac.uk wrote:
are you 'HUP'ing the server at those times to reload the config? I too have noticed change in behaviour with 1.1.5 regarding HUP of the server. 1.1.4 was fine, but a HUP on 1.1.5 can kill the process
1.1.4 was fine by accident... see bugs.freeradius.org. A HUP can still kill 1.1.4. As for why 1.1.5 is more problematic, I have no clue. Alan DeKok. -- http://deployingradius.com - The web site of the book http://deployingradius.com/blog/ - The blog
participants (3)
-
A.L.M.Buxey＠lboro.ac.uk -
Alan DeKok -
Dario Palmisano