I have call recording set to automatically record my extension. I have noticed that when turned on and I am in a call, I have a few more milliseconds of delay between the end of my sentence and the beginning of the caller’s sentence. The delay is noticeable and causes us to “walk” on each other at times. If I turn call recording off the issue goes away.
Any insight on how to continue to use call recording without that delay would be appreciated.
My guess is that you need to add some “oomph” to your machine to handle the extra load.
hmmm… oomph?
This same issue is even worse if I forward a call to a cell phone. For example: If my cell phone # is in the ring group and I answer the cell, the “walking on each other” is even worse.
how much oomph do you have now? . . .
Recording takes cpu cycles, and some memory, both are compounded if you need to transcode, So most folks have enough oomph, but if you are trying to do it on an underpowered system like a raspberry or such, then you won’t have enough oompth.
CPU:
4 vCore
RAM:
8 MB
Storage:
25 GB SSD
Low volume server - Only 1 or 2 calls at a time are ever going through it
If local, you have enough oomph. if the machine is virtual, then “who knows?” , is it a real machine or virtual, if virtual is it KVM or just a container?
vm - Cloud hosted at Vultr
Sorry my experience doesn’t reflect that, I have dozens of PBI right there on Vultr , most are I g , 1 CPU ( the 5 buck chuck) with as many as 30 users with no such trouble (hylafaxing included ;- ), more than enough oomph on debian, can’t speak for other systems.
Just thinking . . . maybe try using just one CPU, I have always been suspicious of Asterisk’s multi-threading scheduling and you are really just wasting money there.
I used Centos 7 - What are your thoughts on Centos 7 versus debian?
It’s a religious war that has no winners.
Anecdotally I have machines that were once edge and are now stretch, they went from asterisk 1.4 to 14 , asterisk@home to FreePBX 13, occasionally restored but never rebuilt. The seed machine has been cloned and recloned , redeployed and re-re deployed. software added and occasionally purged. Is that an implied answer?
I could add, that there has rarely been a problem with missing drivers when redeploying on “hardware” although I rarely use “hardware” anymore (NUC’s being an exception, for a few months until the debian kernel “Caught up”), good tools are Mondorescue and Clonezilla, Mondo cos it works and Clonezilla because it is embedded in archlinux , that is likely a boot choice in many cloud services.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.