16 Aug
2023
16 Aug
'23
2:55 p.m.
Great feedback. Thanks.
Do we have all of this covered in the docs? That we’re using PUSH over UDP in many cases,
so things can get out of sync
if we have network downtime (as an example).
/O
On 16 Aug 2023, at 12:45, Fred Posner
<fred(a)pgpx.io> wrote:
A restart (hard down) would resync assuming your config is set to initially sync.
DMQ is push oriented, not pull. So a soft down, for example a brief connectivity issue,
will result in the node missing the pushes.
It’s not so much a bug, just the way of design. Some other DMQ systems also work
similarly and the burden/responsibility is notI placed on the sender to “own” that a
member receives and processed the message.
This can be done periodically as a safety belt and or when a node notices a connectivity
issue may have occurred.
It can also not be done for scenarios when it’s not necessary for nodes to be exactly in
sync all the time.
-- Fred Posner
Sent from mobile
Phone: +1 (352) 664-3733
On Aug 16, 2023, at 3:03 AM, Olle E. Johansson
<oej(a)edvina.net> wrote:
So if a node is down and comes back up, do you want to force the sync at startup?
Or is this a safetybelt you want cron to run regurlarly? If so, that sounds like a bug…
Sync is a hard problem.
/O
_______________________________________________
Kamailio (SER) - Development Mailing List
To unsubscribe send an email to sr-dev-leave(a)lists.kamailio.org On 15 Aug 2023, at 23:08, Alex Balashov
<abalashov(a)evaristesys.com> wrote:
Well, if I understand Olle's argument correctly, the reason you want this is because
it is not, in fact, "magically happening in the background", and perhaps if it
were, that would be an acceptable fix as well. ;-)
_______________________________________________
Kamailio (SER) - Development Mailing List
To unsubscribe send an email to sr-dev-leave(a)lists.kamailio.org > On Aug 15, 2023, at 5:04 PM, Fred Posner
<fred(a)pgpx.io> wrote:
There are many scenarios I’ve encountered where I wanted this ability…
One would be a temporary loss of connectivity on a DMQ member node, which results in its
htable being out of sync with the other member nodes.
Another would be an accidental local flush.
More scenarios would be applicable as well.
—fred
> On Aug 15, 2023, at 8:21 AM, Olle E. Johansson <oej(a)edvina.net> wrote:
>
>
>
>> On 15 Aug 2023, at 14:59, Daniel-Constantin Mierla <miconda(a)gmail.com>
wrote:
>>
>> htable: docs for rpc htable.dmqsync
>
>
> When do you need to run this and why? Just curious :-)
>
> I thought it was happening magically in the background.
>
> /O
> _______________________________________________
> Kamailio (SER) - Development Mailing List
> To unsubscribe send an email to sr-dev-leave(a)lists.kamailio.org
_______________________________________________
Kamailio (SER) - Development Mailing List
To unsubscribe send an email to sr-dev-leave(a)lists.kamailio.org
--
Alex Balashov
Principal Consultant
Evariste Systems LLC
Web: https://evaristesys.com
Tel: +1-706-510-6800
_______________________________________________
Kamailio (SER) - Development Mailing List
To unsubscribe send an email to sr-dev-leave(a)lists.kamailio.org