I encountered a behavior with DMQ using excessive CPU when running kamailio 5.8 on AWS Graviton processors (running in Docker on Amazon Linux 2023), even on an otherwise idle system. Finding the issue below pointed me towards using polling. I set the value for worker_usleep to 1000 as a test, and that seems to have resolved the CPU issue, but I have no idea of what a "good" value is for this setting, and 1000 was simply the example present in the documentation. My specific use case for DMQ is synchronizing USRLOC data for ~7,500 users, registering with expiration times of 60 - 300 seconds with 7 - 9 nodes on the DMQ bus. Any general recommendations on tuning values for this setting?
https://github.com/kamailio/kamailio/issues/822
Regards, Kaufman
Just checking to see if anyone had any ideas about tuning on this feature.
From: Ben Kaufman via sr-users sr-users@lists.kamailio.org Sent: Thursday, February 1, 2024 8:38 AM To: Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org Cc: Ben Kaufman bkaufman@bcmone.com Subject: [SR-Users] High CPU with DMQ on ARM64 (graviton) - Tuning values for DMQ worker_usleep
CAUTION: This email originated from outside the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.
I encountered a behavior with DMQ using excessive CPU when running kamailio 5.8 on AWS Graviton processors (running in Docker on Amazon Linux 2023), even on an otherwise idle system. Finding the issue below pointed me towards using polling. I set the value for worker_usleep to 1000 as a test, and that seems to have resolved the CPU issue, but I have no idea of what a "good" value is for this setting, and 1000 was simply the example present in the documentation. My specific use case for DMQ is synchronizing USRLOC data for ~7,500 users, registering with expiration times of 60 - 300 seconds with 7 - 9 nodes on the DMQ bus. Any general recommendations on tuning values for this setting?
https://github.com/kamailio/kamailio/issues/822
Regards, Kaufman
What is it synchronizing with? Bear in mind you can’t mix Kamailio versions when doing DMQ
Regards,
David Villasmil email: david.villasmil.work@gmail.com phone: +34669448337
On Fri, 9 Feb 2024 at 17:17, Ben Kaufman via sr-users < sr-users@lists.kamailio.org> wrote:
Just checking to see if anyone had any ideas about tuning on this feature.
*From:* Ben Kaufman via sr-users sr-users@lists.kamailio.org *Sent:* Thursday, February 1, 2024 8:38 AM *To:* Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org *Cc:* Ben Kaufman bkaufman@bcmone.com *Subject:* [SR-Users] High CPU with DMQ on ARM64 (graviton) - Tuning values for DMQ worker_usleep
*CAUTION:* This email originated from outside the organization. *Do not click links or open attachments* unless you recognize the sender and know the content is safe.
I encountered a behavior with DMQ using excessive CPU when running kamailio 5.8 on AWS Graviton processors (running in Docker on Amazon Linux 2023), even on an otherwise idle system. Finding the issue below pointed me towards using polling. I set the value for worker_usleep to 1000 as a test, and that seems to have resolved the CPU issue, but I have no idea of what a “good” value is for this setting, and 1000 was simply the example present in the documentation. My specific use case for DMQ is synchronizing USRLOC data for ~7,500 users, registering with expiration times of 60 – 300 seconds with 7 – 9 nodes on the DMQ bus. Any general recommendations on tuning values for this setting?
https://github.com/kamailio/kamailio/issues/822
Regards,
Kaufman __________________________________________________________ Kamailio - Users Mailing List - Non Commercial Discussions To unsubscribe send an email to sr-users-leave@lists.kamailio.org Important: keep the mailing list in the recipients, do not reply only to the sender! Edit mailing list options or unsubscribe:
For real-usage it's all the same versions. But, I get the behavior even in synchronizing with nothing - making it the only pool member. It seems to be consistent with the behavior in the mailing list link in my original post. Changing from locking to polling (at 1000us) has dropped the CPU utilization to a normal level, but I don't have any idea if 1000us is a "good" value - it's just the value that's in the example in the documentation.
Kaufman Senior Voice Engineer
E: bkaufman@bcmone.com
SIP.US Client Support: 800.566.9810 | SIPTRUNK Client Support: 800.250.6510 | Flowroute Client Support: 855.356.9768 [img]https://www.sip.us/ [img]https://www.siptrunk.com/ [img]https://www.flowroute.com/
From: David Villasmil david.villasmil.work@gmail.com Sent: Friday, February 9, 2024 10:28 AM To: Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org Cc: Ben Kaufman bkaufman@bcmone.com Subject: Re: [SR-Users] Re: High CPU with DMQ on ARM64 (graviton) - Tuning values for DMQ worker_usleep
CAUTION: This email originated from outside the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.
What is it synchronizing with? Bear in mind you can't mix Kamailio versions when doing DMQ
Regards,
David Villasmil email: david.villasmil.work@gmail.commailto:david.villasmil.work@gmail.com phone: +34669448337
On Fri, 9 Feb 2024 at 17:17, Ben Kaufman via sr-users <sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org> wrote: Just checking to see if anyone had any ideas about tuning on this feature.
From: Ben Kaufman via sr-users <sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org> Sent: Thursday, February 1, 2024 8:38 AM To: Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org> Cc: Ben Kaufman <bkaufman@bcmone.commailto:bkaufman@bcmone.com> Subject: [SR-Users] High CPU with DMQ on ARM64 (graviton) - Tuning values for DMQ worker_usleep
CAUTION: This email originated from outside the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.
I encountered a behavior with DMQ using excessive CPU when running kamailio 5.8 on AWS Graviton processors (running in Docker on Amazon Linux 2023), even on an otherwise idle system. Finding the issue below pointed me towards using polling. I set the value for worker_usleep to 1000 as a test, and that seems to have resolved the CPU issue, but I have no idea of what a "good" value is for this setting, and 1000 was simply the example present in the documentation. My specific use case for DMQ is synchronizing USRLOC data for ~7,500 users, registering with expiration times of 60 - 300 seconds with 7 - 9 nodes on the DMQ bus. Any general recommendations on tuning values for this setting?
https://github.com/kamailio/kamailio/issues/822
Regards, Kaufman __________________________________________________________ Kamailio - Users Mailing List - Non Commercial Discussions To unsubscribe send an email to sr-users-leave@lists.kamailio.orgmailto:sr-users-leave@lists.kamailio.org Important: keep the mailing list in the recipients, do not reply only to the sender! Edit mailing list options or unsubscribe:
-
Ben Kaufman -
David Villasmil