[asterisk-bugs] [JIRA] (ASTERISK-28290) res_resolver_unbound.so: Failed to perform async DNS resolution

[Mark (JIRA)]

     [ https://issues.asterisk.org/jira/browse/ASTERISK-28290?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Mark updated ASTERISK-28290:
----------------------------

    Description: 
After upgrading asterisk 13.24.1 -> 16.1.1, i'm occasionally getting the following error which leads to asterisk not being able to register/re-register with my SIP provider.

{quote}ERROR[-1]:res_resolver_unbound.c:306 unbound_resolver_resolve: Failed to perform async DNS resolution of 'tel.t-online.de'{quote}

As soon, as this message appears,  CPU usage is going up to nearly 100%

As i said, there's no real timeframe when it happens.
- multiple times in a row directly after restarting asterisk
- after few minutes, so initial registration works, re-registration fails
- after hours, so initial registration works, re-registration fails

I can fully circumvent this error with asterisk running perfectly stable when unloading "res_resolver_unbound.so" via module.conf. I think this leads to asterisk using the system resolver? I've attached the console log and what i wonder about is, that although the res_resolver_unbound module is loaded during that test, i don't seen any resolver_unbound DEBUG messages here...only pjsip_resolver. Is that correct behaviour?

  was:
After upgrading asterisk 13.24.1 -> 16.1.1, i'm occasionally getting the following error which leads to asterisk not being able to register/re-register with my SIP provider.

{quote}ERROR[-1]:res_resolver_unbound.c:306 unbound_resolver_resolve: Failed to perform async DNS resolution of 'tel.t-online.de'{quote}

As soon, as this message appears,  CPU usage is going up to nearly 100%

As i said, there's no real timeframe when it happens.
- multiple times in a row directly after restarting asterisk
- after few minutes, so initial registration works, re-registration fails
- after hours, so initial registration works, re-registration fails

I can fully circumvent this error with asterisk running perfectly stable over days when unloading "res_resolver_unbound.so" via module.conf. I think this leads to asterisk using the system resolver? I've attached the console log and what i wonder about is, that although the res_resolver_unbound module is loaded during that test, i don't seen any resolver_unbound DEBUG messages here...only pjsip_resolver. Is that correct behaviour?


> res_resolver_unbound.so: Failed to perform async DNS resolution
> ---------------------------------------------------------------
>
>                 Key: ASTERISK-28290
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-28290
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: Resources/res_resolver_unbound
>    Affects Versions: 16.1.1
>         Environment: OpenBSD 6.4 -current (amd64)
>            Reporter: Mark
>            Severity: Minor
>              Labels: pjsip
>         Attachments: CLI.log
>
>
> After upgrading asterisk 13.24.1 -> 16.1.1, i'm occasionally getting the following error which leads to asterisk not being able to register/re-register with my SIP provider.
> {quote}ERROR[-1]:res_resolver_unbound.c:306 unbound_resolver_resolve: Failed to perform async DNS resolution of 'tel.t-online.de'{quote}
> As soon, as this message appears,  CPU usage is going up to nearly 100%
> As i said, there's no real timeframe when it happens.
> - multiple times in a row directly after restarting asterisk
> - after few minutes, so initial registration works, re-registration fails
> - after hours, so initial registration works, re-registration fails
> I can fully circumvent this error with asterisk running perfectly stable when unloading "res_resolver_unbound.so" via module.conf. I think this leads to asterisk using the system resolver? I've attached the console log and what i wonder about is, that although the res_resolver_unbound module is loaded during that test, i don't seen any resolver_unbound DEBUG messages here...only pjsip_resolver. Is that correct behaviour?



--
This message was sent by Atlassian JIRA
(v6.2#6252)