Re: FreeRADIUS + Postgresql dies unexpectedly
Hi, >> I'd suggest changing sql_query() function in sql_postgresql.c to: >> >> ... >> if (!errormsg) return -1; >> ... >> >> instead of the current block of code { errormsg = "FATAL ERROR" } >> Well I tried this change, you can see the gdb output at: http://dl.dropbox.com/u/797051/radiusd/gdb-radiusd-3.log Hope it helps. > As Duarte is so able to reproduce this bug I guess this is over to him? > As for reproducing this, it's pretty straightforward if I follow the steps I outlined in my first e-mail, as anyone else reproduced this successfully? cheers, Duarte - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Alan DeKok wrote: > >> This is a different implosion this time round to your last one. Looks >> like Alan's SQL patch simply postponed the SIGSEGV to a different part >> of the code. Not sure what should be done, but it looks like errorcode >> also needs to be changed (probably set to NULL explicitly or something). >> Alternatively it might be that the error code table in sql_postgresql.h >> needs updating? No real idea here, Alan? > > I'd suggest changing sql_query() function in sql_postgresql.c to: > >... >if (!errormsg) return -1; >... > > instead of the current block of code { errormsg = "FATAL ERROR" } > As Duarte is so able to reproduce this bug I guess this is over to him? Cheers -- Alexander Clouter .sigmonster says: Cure the disease and kill the patient. -- Francis Bacon - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Alexander Clouter wrote: > This is a different implosion this time round to your last one. Looks > like Alan's SQL patch simply postponed the SIGSEGV to a different part > of the code. Not sure what should be done, but it looks like errorcode > also needs to be changed (probably set to NULL explicitly or something). > Alternatively it might be that the error code table in sql_postgresql.h > needs updating? No real idea here, Alan? I'd suggest changing sql_query() function in sql_postgresql.c to: ... if (!errormsg) return -1; ... instead of the current block of code { errormsg = "FATAL ERROR" } > Although FreeRADIUS should not be dying, it does make me wonder why you > can reproduce this problem so regularly :) It's not FreeRADIUS... it's the PostgreSQL libraries. They are supposed to return an error code and message... they don't. Then, when they don't, they go crazy. We've got to work around bugs in libraries we use. Alan DeKok. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Duarte Fonseca wrote: > > 2009/11/12 Alexander Clouter : > >> You should also compile the whole thing with optimisations turned off >> and debugging symbols in there; you are not doing the former so it might >> make it more difficult to work out what is wrong: >> >> CFLAGS='-O0 -g' ./configure >> > > Here is the gdb output and valgrind with these flags set. > > http://dl.dropbox.com/u/797051/radiusd/gdb-radiusd-2.log > http://dl.dropbox.com/u/797051/radiusd/valgrind-radiusd-2.log > This is a different implosion this time round to your last one. Looks like Alan's SQL patch simply postponed the SIGSEGV to a different part of the code. Not sure what should be done, but it looks like errorcode also needs to be changed (probably set to NULL explicitly or something). Alternatively it might be that the error code table in sql_postgresql.h needs updating? No real idea here, Alan? I only was able to trigger this bug once, and that was during a spanning-tree going crazy (scheduled test of a reboot of one of our core routers) when FreeRADIUS would have been unable to talk to the SQL server. Although FreeRADIUS should not be dying, it does make me wonder why you can reproduce this problem so regularly :) Cheers -- Alexander Clouter .sigmonster says: Feminists just want the human race to be a tie. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Hi Alex, Again thanks for the help. 2009/11/12 Alexander Clouter : > You should also compile the whole thing with optimisations turned off > and debugging symbols in there; you are not doing the former so it might > make it more difficult to work out what is wrong: > > CFLAGS='-O0 -g' ./configure > > Here is the gdb output and valgrind with these flags set. http://dl.dropbox.com/u/797051/radiusd/gdb-radiusd-2.log http://dl.dropbox.com/u/797051/radiusd/valgrind-radiusd-2.log Thank you, Duarte - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Hi John, As long as the hammer does the job. At this stage although I'm aware that RPM packaging is much more powerful my lack of knowledge about it doesn't allow for a more sensible approach. Hopefully this will change time ;) thanks, Duarte 2009/11/12 John Dennis : > On 11/12/2009 12:28 PM, Duarte Fonseca wrote: >> >> Hi John, >> >> The only change I did to the spec file was change the release (and >> later add the --enable-developer) >> >> What I did do was go to the SOURCES dir and unpack the >> freeradius-server archive, change the file in question, pack up >> freeradius-server replacing the original tar.bz2 archive. >> >> I did this assuming rpm build would build using whatever code is in >> that tar.bz2 file. Is this another case of assumption being the mother >> of all f*ckups? > > You can do it this way, but it's like trying to use a hammer to drive a > screw :-) The whole rationale of RPM packaging is based around applying > patches to upstream sources, which in this instance is exactly what you're > doing. > -- > John Dennis > > Looking to carve out IT costs? > www.redhat.com/carveoutcosts/ > - > List info/subscribe/unsubscribe? See > http://www.freeradius.org/list/users.html > - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Duarte Fonseca wrote: > > So the problem seems to still be present, this is how I'm testing, > please let me know if I'm doing something wrong. > > I got the redhat SRPM from > > http://people.redhat.com/jdennis/freeradius-rhel-centos/src/freeradius2-2.1.7-2.el5.src.rpm > > Applied the change to the source file and built the binary rpm, > installed it and re run my test several times, and it kept failing. > git clone http://git.freeradius.org/freeradius-server.git cd freeradius-server git checkout origin/stable git checkout -b mybranch That will give you what is to become 2.1.8 and the patches in there are considered stable. If you want to go with 2.1.7 then: git checkout release_2_1_7 git checkout -b mybranch git cherry-pick 45877bf44b02d418b6fb263a39e5de07ced58b6e You should also compile the whole thing with optimisations turned off and debugging symbols in there; you are not doing the former so it might make it more difficult to work out what is wrong: CFLAGS='-O0 -g' ./configure You Redhat folk might have a Redhat Way(tm) to do the above though? > I've rebuilt the rpm with --enable-developer and captured the > following with gdb and valgrind. > > http://dl.dropbox.com/u/797051/radiusd/gdb-radiusd.log > http://dl.dropbox.com/u/797051/radiusd/valgrind-radiusd.log > To my untrained eye it looks like the bug in in libpq, unless libpq is being handed a mangled pointer by FreeRADIUS. Looks rather similar to: http://archives.postgresql.org/pgadmin-hackers/2009-01/msg00079.php > I also tried replacing the full source with the latest from > git(stable) but I failed to get it to build a rpm. > Ahwell...I can give you the Debian destructions :) Cheers -- Alexander Clouter .sigmonster says: Goodbye, cool world. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
On 11/12/2009 12:28 PM, Duarte Fonseca wrote: Hi John, The only change I did to the spec file was change the release (and later add the --enable-developer) What I did do was go to the SOURCES dir and unpack the freeradius-server archive, change the file in question, pack up freeradius-server replacing the original tar.bz2 archive. I did this assuming rpm build would build using whatever code is in that tar.bz2 file. Is this another case of assumption being the mother of all f*ckups? You can do it this way, but it's like trying to use a hammer to drive a screw :-) The whole rationale of RPM packaging is based around applying patches to upstream sources, which in this instance is exactly what you're doing. -- John Dennis Looking to carve out IT costs? www.redhat.com/carveoutcosts/ - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Hi John, The only change I did to the spec file was change the release (and later add the --enable-developer) What I did do was go to the SOURCES dir and unpack the freeradius-server archive, change the file in question, pack up freeradius-server replacing the original tar.bz2 archive. I did this assuming rpm build would build using whatever code is in that tar.bz2 file. Is this another case of assumption being the mother of all f*ckups? Thank you, Duarte 2009/11/12 John Dennis : > On 11/12/2009 11:38 AM, Duarte Fonseca wrote: >> >> Hi again, >> >> So the problem seems to still be present, this is how I'm testing, >> please let me know if I'm doing something wrong. >> >> I got the redhat SRPM from >> >> >> http://people.redhat.com/jdennis/freeradius-rhel-centos/src/freeradius2-2.1.7-2.el5.src.rpm >> >> Applied the change to the source file and built the binary rpm, >> installed it and re run my test several times, and it kept failing. > > Let's be clear concerning the statement "Applied the change to the source > file". You must add the patch to the spec file and then build the binary > rpm, you can't just patch a source file. That's because rpmbuild unpacks the > tar file, applies *only* the patches listed in the spec file, and then > builds the result. If you don't modify the spec file to add the patch then > the patch will never get applied. > > -- > John Dennis > > Looking to carve out IT costs? > www.redhat.com/carveoutcosts/ > - > List info/subscribe/unsubscribe? See > http://www.freeradius.org/list/users.html > - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
On 11/12/2009 11:38 AM, Duarte Fonseca wrote: Hi again, So the problem seems to still be present, this is how I'm testing, please let me know if I'm doing something wrong. I got the redhat SRPM from http://people.redhat.com/jdennis/freeradius-rhel-centos/src/freeradius2-2.1.7-2.el5.src.rpm Applied the change to the source file and built the binary rpm, installed it and re run my test several times, and it kept failing. Let's be clear concerning the statement "Applied the change to the source file". You must add the patch to the spec file and then build the binary rpm, you can't just patch a source file. That's because rpmbuild unpacks the tar file, applies *only* the patches listed in the spec file, and then builds the result. If you don't modify the spec file to add the patch then the patch will never get applied. -- John Dennis Looking to carve out IT costs? www.redhat.com/carveoutcosts/ - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Hi again, So the problem seems to still be present, this is how I'm testing, please let me know if I'm doing something wrong. I got the redhat SRPM from http://people.redhat.com/jdennis/freeradius-rhel-centos/src/freeradius2-2.1.7-2.el5.src.rpm Applied the change to the source file and built the binary rpm, installed it and re run my test several times, and it kept failing. I've rebuilt the rpm with --enable-developer and captured the following with gdb and valgrind. http://dl.dropbox.com/u/797051/radiusd/gdb-radiusd.log http://dl.dropbox.com/u/797051/radiusd/valgrind-radiusd.log I also tried replacing the full source with the latest from git(stable) but I failed to get it to build a rpm. Is there any other changes that I need to apply to the 2.1.7 source? Thank you, Duarte 2009/11/12 Duarte Fonseca : > Thanks Alan and Alex, > > I thought it was that one, just wanted to make sure as I gave it a > quick test and it seems the problem is still there. > > I'm doing some more tests to make sure it's not me doing something > silly, I'll be in touch if I keep having problems (hopefully with gdb > and valgrind output). > > Thank you, > > Duarte > > 2009/11/12 Alexander Clouter : >> Duarte Fonseca wrote: >>> >>> If you could point out which specific commit(s) address this issue I >>> would be very grateful. >>> >> One day we will persaude Alan to put something more verbose in the git >> commit logs :) >> >> Until then: >> >> git log -p -n1 src/modules/rlm_sql/drivers/rlm_sql_postgresql/ >> >> >> It was the recent commit 45877bf44b02d418b6fb263a39e5de07ced58b6e. >> >> Cheers >> >> -- >> Alexander Clouter >> .sigmonster says: Would you care to drift aimlessly in my direction? >> >> - >> List info/subscribe/unsubscribe? See >> http://www.freeradius.org/list/users.html >> > - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Thanks Alan and Alex, I thought it was that one, just wanted to make sure as I gave it a quick test and it seems the problem is still there. I'm doing some more tests to make sure it's not me doing something silly, I'll be in touch if I keep having problems (hopefully with gdb and valgrind output). Thank you, Duarte 2009/11/12 Alexander Clouter : > Duarte Fonseca wrote: >> >> If you could point out which specific commit(s) address this issue I >> would be very grateful. >> > One day we will persaude Alan to put something more verbose in the git > commit logs :) > > Until then: > > git log -p -n1 src/modules/rlm_sql/drivers/rlm_sql_postgresql/ > > > It was the recent commit 45877bf44b02d418b6fb263a39e5de07ced58b6e. > > Cheers > > -- > Alexander Clouter > .sigmonster says: Would you care to drift aimlessly in my direction? > > - > List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html > - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Duarte Fonseca wrote: > > If you could point out which specific commit(s) address this issue I > would be very grateful. > One day we will persaude Alan to put something more verbose in the git commit logs :) Until then: git log -p -n1 src/modules/rlm_sql/drivers/rlm_sql_postgresql/ It was the recent commit 45877bf44b02d418b6fb263a39e5de07ced58b6e. Cheers -- Alexander Clouter .sigmonster says: Would you care to drift aimlessly in my direction? - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Duarte Fonseca wrote: > Hi Alan, > > If you could point out which specific commit(s) address this issue I > would be very grateful. $ git log src/modules/rlm_sql/drivers/rlm_sql_postgresql ... 45877bf44b02d418b6fb263a39e5de07ced58b6e Alan DeKok. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Hi Alan, If you could point out which specific commit(s) address this issue I would be very grateful. Thank you, Duarte 2009/11/6 Alan DeKok : > Duarte Fonseca wrote: >> Hi list, >> >> Just checking if anyone has any more suggestions on how I should >> proceed with this. > > A fix is in the "stable" branch of git. > > Alan DeKok. > - > List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html > - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: FreeRADIUS + Postgresql dies unexpectedly
Hi Alan, Thanks for the advice, I've tried the same test with radiusd -f and here are the results Some runs just produce this: # radiusd -f Segmentation fault Others produce a more verbose output, hope it helps: # radiusd -f *** glibc detected *** radiusd: double free or corruption (fasttop): 0x09143220 *** === Backtrace: = /lib/libc.so.6[0x3c50f1] /lib/libc.so.6(cfree+0x90)[0x3c8bc0] /lib/libcrypto.so.6(CRYPTO_free+0x3a)[0x2a7a5a] /lib/libcrypto.so.6(X509_VERIFY_PARAM_free+0x29)[0x26c029] /lib/libssl.so.6(SSL_free+0x64)[0x13b8b4] /usr/lib/libpq.so.4[0xfdc4e9] /usr/lib/libpq.so.4(pqReadData+0x1da)[0xfd46ba] /usr/lib/libpq.so.4(PQgetResult+0xa0)[0xfd3180] /usr/lib/libpq.so.4[0xfd3348] /usr/lib/libpq.so.4(PQexec+0x27)[0xfd35e7] /usr/lib/freeradius/rlm_sql_postgresql.so[0x18e7bf] /usr/lib/freeradius/rlm_sql.so(rlm_sql_query+0x52)[0xa6b2e2] /usr/lib/freeradius/rlm_sql.so[0xa68a96] radiusd(modcall+0x94e)[0x8062dfe] radiusd(indexed_modcall+0xa5)[0x805fa65] radiusd(module_accounting+0x2c)[0x805fd4c] radiusd(rad_accounting+0xab)[0x804df9b] radiusd(radius_handle_request+0x65)[0x806e4d5] radiusd[0x8066065] /lib/libpthread.so.0[0xd1949b] /lib/libc.so.6(clone+0x5e)[0x42d42e] === Memory map: 0011-00153000 r-xp 08:02 6291673/lib/libssl.so.0.9.8e 00153000-00157000 rwxp 00042000 08:02 6291673/lib/libssl.so.0.9.8e 00157000-00183000 r-xp 08:02 8986736/usr/lib/libgssapi_krb5.so.2.2 00183000-00184000 rwxp 0002c000 08:02 8986736/usr/lib/libgssapi_krb5.so.2.2 00184000-00186000 r-xp 08:02 6291658/lib/libcom_err.so.2.1 00186000-00187000 rwxp 1000 08:02 6291658/lib/libcom_err.so.2.1 00187000-00189000 r-xp 08:02 6291546/lib/libkeyutils-1.2.so 00189000-0018a000 rwxp 1000 08:02 6291546/lib/libkeyutils-1.2.so 0018a000-0018c000 r-xp 08:02 9085015 /usr/lib/freeradius/rlm_acct_unique-2.1.7.so 0018c000-0018d000 rwxp 1000 08:02 9085015 /usr/lib/freeradius/rlm_acct_unique-2.1.7.so 0018d000-00191000 r-xp 08:02 9085105 /usr/lib/freeradius/rlm_sql_postgresql-2.1.7.so 00191000-00192000 rwxp 4000 08:02 9085105 /usr/lib/freeradius/rlm_sql_postgresql-2.1.7.so 00194000-0019d000 r-xp 08:02 6291484/lib/libcrypt-2.5.so 0019d000-0019e000 r-xp 8000 08:02 6291484/lib/libcrypt-2.5.so 0019e000-0019f000 rwxp 9000 08:02 6291484/lib/libcrypt-2.5.so 0019f000-001c6000 rwxp 0019f000 00:00 0 001d4000-001d5000 r-xp 001d4000 00:00 0 [vdso] 001d5000-002ff000 r-xp 08:02 6291671/lib/libcrypto.so.0.9.8e 002ff000-00312000 rwxp 00129000 08:02 6291671/lib/libcrypto.so.0.9.8e 00312000-00316000 rwxp 00312000 00:00 0 00316000-00328000 r-xp 08:02 8984623/usr/lib/libz.so.1.2.3 00328000-00329000 rwxp 00011000 08:02 8984623/usr/lib/libz.so.1.2.3 00345000-00358000 r-xp 08:02 6291490/lib/libnsl-2.5.so 00358000-00359000 r-xp 00012000 08:02 6291490/lib/libnsl-2.5.so 00359000-0035a000 rwxp 00013000 08:02 6291490/lib/libnsl-2.5.so 0035a000-0035c000 rwxp 0035a000 00:00 0 0035c000-0049a000 r-xp 08:02 6291480/lib/libc-2.5.so 0049a000-0049c000 r-xp 0013e000 08:02 6291480/lib/libc-2.5.so 0049c000-0049d000 rwxp 0014 08:02 6291480/lib/libc-2.5.so 0049d000-004a rwxp 0049d000 00:00 0 004a-00533000 r-xp 08:02 8986750/usr/lib/libkrb5.so.3.3 00533000-00536000 rwxp 00092000 08:02 8986750/usr/lib/libkrb5.so.3.3 0064-0065a000 r-xp 08:02 6291473/lib/ld-2.5.so 0065a000-0065b000 r-xp 00019000 08:02 6291473/lib/ld-2.5.so 0065b000-0065c000 rwxp 0001a000 08:02 6291473/lib/ld-2.5.so 006b9000-006c8000 r-xp 08:02 6291506/lib/libresolv-2.5.so 006c8000-006c9000 r-xp e000 08:02 6291506/lib/libresolv-2.5.so 006c9000-006ca000 rwxp f000 08:02 6291506/lib/libresolv-2.5.so 006ca000-006cc000 rwxp 006ca000 00:00 0 006f4000-006f6000 r-xp 08:02 9085067 /usr/lib/freeradius/rlm_files-2.1.7.so 006f6000-006f7000 rwxp 1000 08:02 9085067 /usr/lib/freeradius/rlm_files-2.1.7.so 00753000-00769000 r-xp 08:02 6291646/lib/libselinux.so.1 00769000-0076b000 rwxp 00015000 08:02 6291646/lib/libselinux.so.1 00787000-00792000 r-xp 08:02 6291458 Aborted # radiusd -f *** glibc detected *** radiusd: double free or corruption (fasttop): 0x0993eeb0 *** === Backtrace: = /lib/libc.so.6[0x1790f1] /lib/libc.so.6(cfree+0x90)[0x17cbc0] /lib/libcrypto.so.6(CRYPTO_free+0x3a)[0x483a5a] /lib/libcrypto.so.6(X509_VERIFY_PARAM_free+0x29)[0x448029] /lib/libssl.so.6(SSL_free+0x64)[0xede8b4] /usr/lib/libpq.so.4[0x3a14e9] /usr/lib/libpq.so.4(pqReadData+0x1da)[0x3996ba] /usr/lib/libpq.so.4(PQgetResult+0xa0)[0x398180] /usr/lib/libpq.so.4[0x39828c] /usr/lib/freeradius/rlm_sql_postgresql.so[0x3887bf] /usr/lib/freeradius/rlm_sql.so(rlm_sql_query+0x52)[0x9e82e2] /usr/lib/freeradius/rlm_sql.so[0x9e5c3b] radiusd(modcall+0x94e)[0x8062dfe] radiusd(indexed_modcall+0xa5)[0x805fa65] radiusd(
Re: FreeRADIUS + Postgresql dies unexpectedly
Duarte Fonseca wrote: > I've got a freeRadius (v2.1.7) install running on CentOs using > postgresql to store accounting data and have noticed that occasionally > freeRadius seems to die unexpectedly. See doc/bugs. You can run the server in foreground mode (radiusd -f), too. If it dies after a few packets, the problem should be pretty simple to find && fix. Alan DeKok. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
FreeRADIUS + Postgresql dies unexpectedly
Hi, I've got a freeRadius (v2.1.7) install running on CentOs using postgresql to store accounting data and have noticed that occasionally freeRadius seems to die unexpectedly. After some time investigating this and going over the logs I can now reproduce this behavior easily by following these steps: 1 - In my sql configuration file (/etc/raddb/postgresql.conf) I specify server = "localhost" (freeRADIUS connects to the database through TCP/IP) 2 - Restart radiusd 3 - Use radclient to send several accounting requests without an Acct-session-Time attribute (radclient -p 200 -f 1.txt 10.1.60.141:1813 acct secret) After a few requests radiusd dies. If i change my sql configuration file (/etc/raddb/postgresql.conf) so server = "" (freeRADIUS connects to DB using Unix-domain socket) radiusd doesn't die. If i run radiusd in debug mode it doesn't die. If i send accounting requests with the Acct-Session-Time set to a positive integer radiusd doesn't die. The following are extracts from the freeRADIUS log and Postgresql log during a test where the radiusd dies extract from radius.log # cat /var/log/radius/radius.log Thu Oct 22 16:20:30 2009 : Info: rlm_sql (sql): Driver rlm_sql_postgresql (module rlm_sql_postgresql) loaded and linked Thu Oct 22 16:20:30 2009 : Info: rlm_sql (sql): Attempting to connect to postg...@localhost:/gapdb Thu Oct 22 16:20:30 2009 : Info: Loaded virtual server Thu Oct 22 16:20:30 2009 : Info: Ready to process requests. Thu Oct 22 16:20:51 2009 : Error: [sql] stop packet with zero session length. [user '0017f24eff31', nas '172.20.200.66'] Thu Oct 22 16:20:51 2009 : Error: [sql] stop packet with zero session length. [user '0012f0aac218', nas '172.20.200.66'] Thu Oct 22 16:20:55 2009 : Error: [sql] stop packet with zero session length. [user 'admin', nas '172.20.200.226'] Thu Oct 22 16:20:55 2009 : Error: [sql] stop packet with zero session length. [user 'JoePublic7', nas '172.20.200.226'] Thu Oct 22 16:20:55 2009 : Error: [sql] stop packet with zero session length. [user 'admin', nas '172.20.200.226'] Thu Oct 22 16:20:55 2009 : Error: [sql] stop packet with zero session length. [user 'admin', nas '172.20.200.226'] extract from postgresql log <2009-10-22 16:20:55.095 BST> LOG: SSL error: sslv3 alert bad record mac <2009-10-22 16:20:55.095 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.095 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.097 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.097 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.097 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.097 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.098 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.098 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.098 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.098 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.098 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.098 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.103 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.103 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.103 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.103 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.103 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.103 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.103 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.104 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.105 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.105 BST> LOG: unexpected EOF on client connection <2009-10-22 16:20:55.111 BST> LOG: could not receive data from client: Connection reset by peer <2009-10-22 16:20:55.111 BST> LOG: unexpected EOF on client connection # radiusd -v radiusd: FreeRADIUS Version 2.1.7, for host i386-redhat-linux-gnu, built on Sep 18 2009 at 10:59:17 How would you guys advise me to proceed with resolving this issue, is it a bug, is it something wrong in my configuration? Thank you, Duarte -- Duarte Fonseca -- Mobile: +44 753 4262674 E-mail: fonseca.dua...@gmail.com - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html