[Devel] 477 on dispatcher
Helge Waastad
helge at smartnet.no
Wed Feb 7 15:04:34 CET 2007
Hi,
what I currently see is that there might be some timer issues for me.
When I have not enabled disable_dns_blacklists=yes, many of the messages
received from the UAC's are retransmitted to the proxies, even if I
generate a 100 response on the proxy back to the LB to stop
retransmissions.
....And since I also are trying the presence module, a lot of messages
are retransmitted.
For now, I have to disable blacklists.
br hw
On ons, 2007-02-07 at 15:00 +0200, Bogdan-Andrei Iancu wrote:
> Hi Helge,
>
> looks like the 477 is indeed triggered by some entries in the temporary
> blacklist of the DNS resolver. Most probably this entries are generated
> because of a 408 timeout with no reply. can you check on this? Probably
> is also important for you to know that you have such situation on your
> LBs :).
>
> For example, on LB1, the address of the SP1 is timeout-ed.
>
> regards,
> bogdan
>
> Helge Waastad wrote:
> > Hi,
> > yes I got an 477 and fifo list_blacklists on dispatcher layer shows:
> > (I have LB and SP running on two hosts, LB on 5060, SP on 5080)
> > Having #disable_dns_blacklists=yes in my config.
> >
> > LB #1 <IP #1>:
> > List:: dns owner=17 flags=6
> > Rule:: flags=0
> > IP:: <IP #1>
> > Mask:: 255.255.255.255
> > Proto:: 1
> > Port:: 5080
> > Expire:: 7089
> > Rule:: flags=0
> > IP:: <IP #1>
> > Mask:: 255.255.255.255
> > Proto:: 1
> > Port:: 5060
> > Expire:: 7092
> >
> >
> > LB #2 <IP #2>:
> >
> > List:: dns owner=17 flags=6
> > Rule:: flags=0
> > IP:: <IP #1>
> > Mask:: 255.255.255.255
> > Proto:: 1
> > Port:: 5060
> > Expire:: 7064
> >
> >
> > br hw
> >
> >
> >
> > On ons, 2007-02-07 at 12:06 +0200, Bogdan-Andrei
> > Iancu wrote:
> >
> >> ok - thanks ; that wil be really helpful to see if there is a bug or it
> >> is normal behaviour of blacklists ;)
> >>
> >> regards,
> >> bogdan
> >>
> >> Helge Waastad wrote:
> >>
> >>> Hi,
> >>> I will turn back to the original config and give you feedback on this
> >>> the moment I have managed to force the 477 response
> >>>
> >>> br hw
> >>>
> >>>
> >>>
> >>> On ons, 2007-02-07 at 11:02 +0200, Bogdan-Andrei Iancu wrote:
> >>>
> >>>
> >>>> Hi Helge,
> >>>>
> >>>> I'm trying to go to the bottom of this problem .... was it the case of
> >>>> a local generated 408 timeout with no reply received? Because this case
> >>>> is considered a failure that triggers DNS-based failover (RFC3262) and
> >>>> if disable_dns_blacklist=no, the destination IP address will be
> >>>> blacklisted for 4 minutes.
> >>>>
> >>>> can you check if this was the case?
> >>>>
> >>>> you can inspect the content of the blacklist via MI command
> >>>> "list_blacklists"
> >>>>
> >>>> regards,
> >>>> bogdan
> >>>>
> >>>> Helge Waastad wrote:
> >>>>
> >>>>
> >>>>> Hi,
> >>>>> that helped.
> >>>>>
> >>>>> Thanks a lot.
> >>>>>
> >>>>>
> >>>>> br he
> >>>>>
> >>>>>
> >>>>> On tir, 2007-02-06 at 20:13 +0200, Bogdan-Andrei Iancu wrote:
> >>>>>
> >>>>>
> >>>>>
> >>>>>> Hi Helge,
> >>>>>>
> >>>>>> somebody else reported something similar :
> >>>>>> http://www.openser.org/pipermail/devel/2007-February/005705.html
> >>>>>>
> >>>>>> try to see if:
> >>>>>> disable_dns_blacklist=yes
> >>>>>> has any effect.
> >>>>>>
> >>>>>> regards,
> >>>>>> bogdan
> >>>>>>
> >>>>>>
> >>>>>> Helge Waastad wrote:
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>> Hi,
> >>>>>>> I just want to see if anyone else has seen the same as I have.
> >>>>>>>
> >>>>>>> Every now and then I get a 477 Unfortunately error on sending to next
> >>>>>>> hop occurred (477/TM)
> >>>>>>>
> >>>>>>> on my stateful dispatcher.
> >>>>>>>
> >>>>>>> I'm basically using the same config as I have been using for my 1.1.1
> >>>>>>> setup.
> >>>>>>>
> >>>>>>> I will try to debug further
> >>>>>>>
> >>>>>>> br hw
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>> _______________________________________________
> >>>>>>> Devel mailing list
> >>>>>>> Devel at openser.org
> >>>>>>> http://openser.org/cgi-bin/mailman/listinfo/devel
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>
> >>>>>
> >>>>>
> >>>
> >>>
> >
> >
>
More information about the Devel
mailing list