random SIGSEGV after upgrade to 3.0.20
Hi, we upgraded our installation to 3.0.20 last week and started experiencing random SIGSEGV crashes on both of our radius servers. Appearently this is related to our eduroam setup. After adding 'status_check = status-server' to the eduroam home_server blocks the crashes stopped. We are unsure whether this is a bug, but it is sort of puzzling. Did anybody else experience something similar? Best regards, Ferdinand -- Ferdinand Goldmann System Administrator Information Management JOHANNES KEPLER UNIVERSITY LINZ Altenberger Straße 69 Hochschulfond Building, HF9902 4040 Linz, Austria P +43 732 2468 3925 ferdinand.goldmann@jku.at www.jku.at/im
On Mon, 2020-03-02 at 10:48 +0100, Ferdinand Goldmann wrote:
we upgraded our installation to 3.0.20 last week
What version were you running before?
and started experiencing random SIGSEGV crashes on both of our radius servers. Appearently this is related to our eduroam setup. After adding 'status_check = status-server' to the eduroam home_server blocks the crashes stopped.
Could you try the latest v3.0.x HEAD? There's at least one fix that went in which I think could be related. If you need packages then you can get them from https://packages.networkradius.com/ - change the repo URLs from /releases/ to /continuous/ to get the most recent (i.e. pre-3.0.21) version. It would be good to confirm if this is fixed before 3.0.21 is released. -- Matthew
Hi! On Mon, 2 Mar 2020, Matthew Newton wrote:
On Mon, 2020-03-02 at 10:48 +0100, Ferdinand Goldmann wrote:
we upgraded our installation to 3.0.20 last week
What version were you running before?
3.0.17
Could you try the latest v3.0.x HEAD? There's at least one fix that went in which I think could be related.
If you need packages then you can get them from https://packages.networkradius.com/ - change the repo URLs from /releases/ to /continuous/ to get the most recent (i.e. pre-3.0.21) version.
It would be good to confirm if this is fixed before 3.0.21 is released.
I will check this out and report on how it's working! Regards Ferdinand -- Ferdinand Goldmann System Administrator Information Management JOHANNES KEPLER UNIVERSITY LINZ Altenberger Straße 69 Hochschulfond Building, HF9902 4040 Linz, Austria P +43 732 2468 3925 ferdinand.goldmann@jku.at www.jku.at/im
On Mon, 2 Mar 2020, Matthew Newton wrote:
Could you try the latest v3.0.x HEAD? There's at least one fix that went in which I think could be related.
I compiled the latest v3 branch version: # /opt/sbin/radiusd -v radiusd: FreeRADIUS Version 3.0.21 (git #395a247c4), for host x86_64-unknown-freebsd11.3, built on Mar 2 2020 at 14:51:03 It has been running on one production server since yesterday without any problems whatsoever, roughly +24 hours by now. I deliberately removed the 'status_check = status-server' statements from the configuration which we used to mitigate the previous crashes. Since after the switch to 3.0.20 both production servers started crashing within hours while this one has now been running for over a day without a crash it seems the bug is fixed. Regards, Ferdinand -- Ferdinand Goldmann System Administrator Information Management JOHANNES KEPLER UNIVERSITY LINZ Altenberger Straße 69 Hochschulfond Building, HF9902 4040 Linz, Austria P +43 732 2468 3925 ferdinand.goldmann@jku.at www.jku.at/im
On Tue, 2020-03-03 at 16:11 +0100, Ferdinand Goldmann wrote:
On Mon, 2 Mar 2020, Matthew Newton wrote: Since after the switch to 3.0.20 both production servers started crashing within hours while this one has now been running for over a day without a crash it seems the bug is fixed.
Sounds good, thanks. -- Matthew
participants (2)
-
Ferdinand Goldmann -
Matthew Newton