Error: WARNING: Unresponsive child for request in module sql component accounting
Hello all!! I facing this problem with my Freeradius 2.1.3, and I don't know how to solve it :( My NAS is sending only accounting registers to my freeradius server. My freeradius server, is configured to store these registers into a MySQL server. I have configured "max_request_time = 120", in the case of MySQL slow performance, but the problem perssits. My questions are: - how can I solve this issue? - When a child is unresponsive, the register is lost, or another child stores it into the database? Thx in advance. Regards RADIUSD.CONF: prefix = /usr/local exec_prefix = ${prefix} sysconfdir = ${prefix}/etc localstatedir = ${prefix}/var sbindir = ${exec_prefix}/sbin logdir = ${localstatedir}/log/radius raddbdir = ${sysconfdir}/raddb radacctdir = ${logdir}/radacct name = radiusd confdir = ${raddbdir} run_dir = ${localstatedir}/run/radiusd db_dir = ${raddbdir} libdir = ${exec_prefix}/lib pidfile = ${run_dir}/${name}.pid max_request_time = 120 cleanup_delay = 1 max_requests = 10240 listen { type = auth ipaddr = * port = 0 } listen { ipaddr = * port = 0 type = acct } hostname_lookups = no allow_core_dumps = no regular_expressions = yes extended_expressions = yes log { destination = files file = ${logdir}/radius.log syslog_facility = daemon stripped_names = no auth = no auth_badpass = no auth_goodpass = no } checkrad = ${sbindir}/checkrad security { max_attributes = 200 reject_delay = 1 status_server = yes } proxy_requests = yes $INCLUDE proxy.conf $INCLUDE clients.conf thread pool { start_servers = 50 max_servers = 100 min_spare_servers = 10 max_spare_servers = 30 max_requests_per_server = 5 } modules { $INCLUDE ${confdir}/modules/ $INCLUDE eap.conf $INCLUDE sql.conf $INCLUDE sql/mysql/counter.conf } instantiate { exec expr expiration logintime } $INCLUDE policy.conf $INCLUDE sites-enabled/
I facing this problem with my Freeradius 2.1.3, and I don't know how to solve it :(
My NAS is sending only accounting registers to my freeradius server. My freeradius server, is configured to store these registers into a MySQL server. I have configured "max_request_time = 120", in the case of MySQL slow performance, but the problem perssits.
No, you don't have a problem with radius server but with sql one. Perhaps you should look into the server that does have a problem (sql) and not the one that doesn't (radius). There is nothing radius server can tell you about why are sql queries running slow. Ivan Kalik Kalik Informatika ISP
Thx Ivan, and do you know if the accouting registers is lost? or another child retries the insert into the database? thx Regards On Miércoles 25 Febrero 2009 14:09:44 tnt@kalik.net wrote:
I facing this problem with my Freeradius 2.1.3, and I don't know how to solve it :(
My NAS is sending only accounting registers to my freeradius server. My freeradius server, is configured to store these registers into a MySQL server. I have configured "max_request_time = 120", in the case of MySQL slow performance, but the problem perssits.
No, you don't have a problem with radius server but with sql one. Perhaps you should look into the server that does have a problem (sql) and not the one that doesn't (radius). There is nothing radius server can tell you about why are sql queries running slow.
Ivan Kalik Kalik Informatika ISP
- List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
and do you know if the accouting registers is lost? or another child retries the insert into the database?
They usually are - there are no handles to write to the database as the whole server gets blocked. I haven't seen tha case where single handle would dia and the rest of them would continue working. This is usually terminal state of radius-sql server connection problem. Ivan Kalik Kalik Informatika ISP
HI again Ivan Do you know how can I modify the source code, in order to print in the error message the exact time when the child was created/forked? If I want to find errors in my sql backend server, I need to know the exact time when the requests are lost...... checking the code, It looks like this error message is printed only when the server detects threads that are stalled..................... Thx again. Regards On Miércoles 25 Febrero 2009 14:09:44 tnt@kalik.net wrote:
I facing this problem with my Freeradius 2.1.3, and I don't know how to solve it :(
My NAS is sending only accounting registers to my freeradius server. My freeradius server, is configured to store these registers into a MySQL server. I have configured "max_request_time = 120", in the case of MySQL slow performance, but the problem perssits.
No, you don't have a problem with radius server but with sql one. Perhaps you should look into the server that does have a problem (sql) and not the one that doesn't (radius). There is nothing radius server can tell you about why are sql queries running slow.
Ivan Kalik Kalik Informatika ISP
- List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
First, optimize your database, add indexes and use storage engines which will give you the best performance... Increase interim-update interval After that, try to add more handlers and see what will happen. If you're using mysql, check tuning-primer.sh and mysqltuner.pl scripts. On Thu, Feb 26, 2009 at 10:23 AM, magicboiz <magicboiz@gmail.com> wrote:
HI again Ivan
Do you know how can I modify the source code, in order to print in the error message the exact time when the child was created/forked? If I want to find errors in my sql backend server, I need to know the exact time when the requests are lost...... checking the code, It looks like this error message is printed only when the server detects threads that are stalled.....................
Thx again. Regards
On Miércoles 25 Febrero 2009 14:09:44 tnt@kalik.net wrote:
I facing this problem with my Freeradius 2.1.3, and I don't know how to solve it :(
My NAS is sending only accounting registers to my freeradius server. My freeradius server, is configured to store these registers into a MySQL server. I have configured "max_request_time = 120", in the case of MySQL slow performance, but the problem perssits.
No, you don't have a problem with radius server but with sql one. Perhaps you should look into the server that does have a problem (sql) and not the one that doesn't (radius). There is nothing radius server can tell you about why are sql queries running slow.
Ivan Kalik Kalik Informatika ISP
- List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
- List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
magicboiz wrote:
Do you know how can I modify the source code, in order to print in the error message the exact time when the child was created/forked? If I want to find errors in my sql backend server, I need to know the exact time when the requests are lost...... checking the code, It looks like this error message is printed only when the server detects threads that are stalled.....................
That message is printed "max_request_time" seconds after the packet is received. The stall happens at some point during that time period. Alan DeKok.
Thx Alan I coded a very simple script to check the mysl server latency: #!/bin/sh ExecQuery () { echo "select count(*) from radacct;" | mysql -h mysqlserver -u radiususer --password=radiuspassword radaccoutingdb } while [ -n 1 ] do echo -n "starting query....." date time ExecQuery echo -n "query finished......" date echo "" echo "" sleep 300 done Executing this script during last 24h (with nohup), from the radius machine, I got always latencies below 1 second........ How can the script get always latencies under 1sec, and the radius threads on teh other hand get latencies over max-request-time? I'm a little bit lost.... :( thx again On Jueves 26 Febrero 2009 11:00:31 Alan DeKok wrote:
magicboiz wrote:
Do you know how can I modify the source code, in order to print in the error message the exact time when the child was created/forked? If I want to find errors in my sql backend server, I need to know the exact time when the requests are lost...... checking the code, It looks like this error message is printed only when the server detects threads that are stalled.....................
That message is printed "max_request_time" seconds after the packet is received. The stall happens at some point during that time period.
Alan DeKok. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Executing this script during last 24h (with nohup), from the radius machine, I got always latencies below 1 second........
Below 1 second or below 1/100th of a second? Near 1 second for such a simple query is *very* slow. Have in mind that insert into radacct is going to take some 100 times longer then a select. Ivan Kalik Kalik Informatika ISP
participants (4)
-
Alan DeKok -
magicboiz -
Marinko Tarlac -
tnt＠kalik.net