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

img

img

 

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

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: