pinger without peers

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

pinger without peers

Matus UHLAR - fantomas
Hello,

does it have sense to run pinger without having cache peers configured?

if I get the "Network DB Statistics:" output properly, it seems that 33% of
hosts is unreachable.

--
Matus UHLAR - fantomas, [hidden email] ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
Linux IS user friendly, it's just selective who its friends are...
_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: pinger without peers

Amos Jeffries
Administrator
On 10/03/2017 6:36 a.m., Matus UHLAR - fantomas wrote:
> Hello,
>
> does it have sense to run pinger without having cache peers configured?
>
> if I get the "Network DB Statistics:" output properly, it seems that 33% of
> hosts is unreachable.
>

The code using it is called peer-selection, but that includes choosing
between origin servers (DIRECT) as well as cache_peer lines. So it can
affect traffic routing even if you dont have cache_peer. Though less
that RTT based cache_peer LB algorithms.

The sad results are due to admin ICMP echo blocking. Hopefully that is
all they are blocking though and not other mandatory ICMP codes like
path-MTU discovery.

Amos

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: pinger without peers

Matus UHLAR - fantomas
>On 10/03/2017 6:36 a.m., Matus UHLAR - fantomas wrote:
>> does it have sense to run pinger without having cache peers configured?
>>
>> if I get the "Network DB Statistics:" output properly, it seems that 33% of
>> hosts is unreachable.

On 10.03.17 08:33, Amos Jeffries wrote:
>The code using it is called peer-selection, but that includes choosing
>between origin servers (DIRECT) as well as cache_peer lines. So it can
>affect traffic routing even if you dont have cache_peer. Though less
>that RTT based cache_peer LB algorithms.

does that mean pinger output is used for servers with multiple addresses?
does it only ping servers with multiple addresses?

>The sad results are due to admin ICMP echo blocking. Hopefully that is
>all they are blocking though and not other mandatory ICMP codes like
>path-MTU discovery.

of course. But since I'm trying to solve different problem, I'm thinking of
trurning off everything that can cause issues.

--
Matus UHLAR - fantomas, [hidden email] ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
"To Boot or not to Boot, that's the question." [WD1270 Caviar]
_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users
Reply | Threaded
Open this post in threaded view
|

Re: pinger without peers

Amos Jeffries
Administrator
On 10/03/2017 10:28 p.m., Matus UHLAR - fantomas wrote:

>> On 10/03/2017 6:36 a.m., Matus UHLAR - fantomas wrote:
>>> does it have sense to run pinger without having cache peers configured?
>>>
>>> if I get the "Network DB Statistics:" output properly, it seems that
>>> 33% of
>>> hosts is unreachable.
>
> On 10.03.17 08:33, Amos Jeffries wrote:
>> The code using it is called peer-selection, but that includes choosing
>> between origin servers (DIRECT) as well as cache_peer lines. So it can
>> affect traffic routing even if you dont have cache_peer. Though less
>> that RTT based cache_peer LB algorithms.
>
> does that mean pinger output is used for servers with multiple addresses?
> does it only ping servers with multiple addresses?

All servers that are selected for use get pinged, but use of the ping
results is uncommon, usually the data comes back far too late to affect
the immediate transaction IME. It is probably of some use on persistent
connections.

>
>> The sad results are due to admin ICMP echo blocking. Hopefully that is
>> all they are blocking though and not other mandatory ICMP codes like
>> path-MTU discovery.
>
> of course. But since I'm trying to solve different problem, I'm thinking of
> trurning off everything that can cause issues.
>

Nod. It is of minor relevance on most Gbit speed (or even Mbit speed)
networks today.

Amos

_______________________________________________
squid-users mailing list
[hidden email]
http://lists.squid-cache.org/listinfo/squid-users