Do you have any server that is timing out? The registrations are handled sequentialy and if a server is timing out, then it will take a long time for opensips to handle all the records in a slot and it may overide with the next scheduled timer task for the next slot. Try to increase the number of slots (hash_slot) and the timer interval parameter.
-ovidiu On Wed, Mar 16, 2016 at 12:48 AM, Alexander Mustafin <[email protected]> wrote: > Hi, Ovidiu. > > Yes, I use FQDN for registrar and somebody already said me here, that is the > possible problem. > > I see next errors in the log: > > Mar 16 03:48:36 node1 opensips[25954]: CRITICAL:core:timer_ticker: timer > handler <uac_reg_check> lasted (2180000 us) for more than timer tick > (1000000 us) -> potential timer shifting > Mar 16 03:48:56 node1 opensips[25954]: CRITICAL:core:timer_ticker: timer > handler <uac_reg_check> lasted (2740000 us) for more than timer tick > (1000000 us) -> potential timer shifting > Mar 16 03:49:36 node1 opensips[25954]: CRITICAL:core:timer_ticker: timer > handler <uac_reg_check> lasted (2670000 us) for more than timer tick > (1000000 us) -> potential timer shifting > > > Do you need gdb of any process of opensips? > > Best regards, > Alexander Mustafin > [email protected] > > > > > 15 марта 2016 г., в 21:47, Ovidiu Sas <[email protected]> написал(а): > > On Tue, Mar 15, 2016 at 8:09 AM, Alexander Mustafin > <[email protected]> wrote: > > Hi there! > > Perhaps I don’t understand logic of this module, but I have strange problem: > > version: opensips 1.11.6-tls (x86_64/linux) > > I have 223 entries in my db. Previously, preferences of module was like > these: > modparam("uac_registrant", "hash_size", 2) > modparam("uac_registrant", "db_url", "SQL_URL") > modparam("uac_registrant", "table_name", "REGISTRANT_TABLE") > modparam("uac_registrant", "timer_interval", 100) > > All works fine, but I had awful CPU load, almost 100%. > > > In this setup, assuming that you had an even distribution, you would > have around 223/4=55 records per hash slot and each slot should be > checked every 100/4=25s. > Can you attach with gdb to the process to see what is happening there? > Do you see any errors in the logs? > > Then, I changed it to > > modparam("uac_registrant", "hash_size", 5) > modparam("uac_registrant", "db_url", "SQL_URL") > modparam("uac_registrant", "table_name", "REGISTRANT_TABLE") > modparam("uac_registrant", "timer_interval", 640) > > Processor feels good now, but I experience problem with registrations: > opensipsctl fifo reg_list |grep AOR|wc -l > 166 > > Opensips shows only 166 entries in memory, so, I have no idea where is my > other registrations. > > How to tie hash_size, timer_interval and value in expiry column? > > See the Overview in the documentation. > The expiry column is stores the binding's expiration time. > > And why I > see this information for some of my numbers, where registration_t_out less > than last_register_sent: > > state:: REGISTERED_STATE > last_register_sent:: Tue Mar 15 06:25:38 2016 > registration_t_out:: Tue Mar 15 06:18:57 2016 > registrar:: sip:some.registrar > binding:: sip:[email protected]:5070 > > > Do you see any errors in the logs? > Are you using FQDNs for registrar? > > > Thanks in advance! > > Best regards, > Alexander Mustafin > [email protected] > > > > > > _______________________________________________ > Users mailing list > [email protected] > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > > > > -- > VoIP Embedded, Inc. > http://www.voipembedded.com > > _______________________________________________ > Users mailing list > [email protected] > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > > > _______________________________________________ > Users mailing list > [email protected] > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > -- VoIP Embedded, Inc. http://www.voipembedded.com _______________________________________________ Users mailing list [email protected] http://lists.opensips.org/cgi-bin/mailman/listinfo/users
