Hi all,
we have a landscape with 2 proxy servers,
only one has the permission to browse the internet through the firewall (i call it parent), the other is at an other location connected through a WAN line which is most time verry busy. ( called son ). Son-proxy forwards all request to parent and does never try direct request:
cache_peer parent.company.com parent 8080 0 no-query default
never_direct allow all
The problem is that sometimes the son-proxy gives error : "Unable to forward this request at this time" indicating that the only parent he has died and the son can not ping the parent, but the parent has not died!
From other systems ( located in the same network as the son ) i can ping the parent-proxy.
Is it possible that the parent-proxy blocks the son-proxy because of some reason? Are there any parameters which do that with client which cause lot of http requests?
Or is it possible that the son proxy has for some reason marked the parent as dead because of some request timeouts, and does not try to connect to it again for a specified time ?
When this effect orly lasts for some time i.e. 30 minutes and then the son-proxy works again ...
I did a lotz of work trying to find any routing or any other network errors when the effect happens, but i were unsuccessful, in my opinion the problem could only cause at the son- or parent-proxy.
both proxies run red-hat Linux 7.3 with squid-2.4.STABLE6-6.7.3
thanks in advance for any hints ...
Mark
we have a landscape with 2 proxy servers,
only one has the permission to browse the internet through the firewall (i call it parent), the other is at an other location connected through a WAN line which is most time verry busy. ( called son ). Son-proxy forwards all request to parent and does never try direct request:
cache_peer parent.company.com parent 8080 0 no-query default
never_direct allow all
The problem is that sometimes the son-proxy gives error : "Unable to forward this request at this time" indicating that the only parent he has died and the son can not ping the parent, but the parent has not died!
From other systems ( located in the same network as the son ) i can ping the parent-proxy.
Is it possible that the parent-proxy blocks the son-proxy because of some reason? Are there any parameters which do that with client which cause lot of http requests?
Or is it possible that the son proxy has for some reason marked the parent as dead because of some request timeouts, and does not try to connect to it again for a specified time ?
When this effect orly lasts for some time i.e. 30 minutes and then the son-proxy works again ...
I did a lotz of work trying to find any routing or any other network errors when the effect happens, but i were unsuccessful, in my opinion the problem could only cause at the son- or parent-proxy.
both proxies run red-hat Linux 7.3 with squid-2.4.STABLE6-6.7.3
thanks in advance for any hints ...
Mark