5

We’re trying to set a bunch of Apaches 2.4.18 to proxy pass the requests it receives to our partner's upstream server. Our partner uses Amazon’s Elastic Load Balancing and thus the only we know about their servers is its DNS names.

The TTL of the DNS records is 60 seconds and I’d like to know if Apache can honour that ttl, keeping the connection alive as long as the DNS record is valid and then requesting the translation when the TTL has expired.

Using mod_proxy DisableReuse = on forces opening a new connection every time a resource is needed upstream. That would do the trick as long as the underneath operating system does the DNS TTL caching. If not, every time a new resource is needed Apache will force a new DNS request, increasing the response time.

I’ve thought of playing with the mod_proxy ttl and timeout parameters, but I think I’m not correctly solving the problem. According to the docs, the mod_proxy’s timeout parameter controls the time a socket will wait for data from upstream, but I’m not sure if the Apache instance will close the connection an open a new one. Also, playing with the timeout is error prone, because a lower value may sent an wrong answer to the client.

I’ve spend a few time trying to tackle this setup with no joy. Is there any special setup to cover that scenario? Or perhaps I’ve skipped something? Any help would be appreciated.

Best,

Gustau

Improve this question

2 Answers 2

Reset to default
5

The only option that can be helpful for your use case is disablereuse=On.

The DNS queries are performed by resolver part of glibc. The results are not cached by this library, or by OS. In apache case the DNS results are cached by apache worker process. You can have a DNS service, like nscd or dnsmasq, that is doing the DNS caching.

Here are excerpts from the apache documentation.

https://httpd.apache.org/docs/2.4/mod/mod_proxy.html#workers

DNS resolution for origin domains

DNS resolution happens when the socket to the origin domain is created for the first time. When connection reuse is enabled, each backend domain is resolved only once per child process, and cached for all further connections until the child is recycled. This information should to be considered while planning DNS maintenance tasks involving backend domains. Please also check ProxyPass parameters for more details about connection reuse.

https://httpd.apache.org/docs/2.4/mod/mod_proxy.html#proxypass

disablereuse (by default is "Off") This parameter should be used when you want to force mod_proxy to immediately close a connection to the backend after being used, and thus, disable its persistent connection and pool for that backend. This helps in various situations where a firewall between Apache httpd and the backend server (regardless of protocol) tends to silently drop connections or when backends themselves may be under round- robin DNS. When connection reuse is enabled each backend domain is resolved (with a DNS query) only once per child process and cached for all further connections until the child is recycled. To disable connection reuse, set this property value to On.

Improve this answer
1
  • Thank you @Mircea Vutcovici. I've played with both nscd and dnsmasq, but our solution won't be deployed in our machines, but in our customer machines instead. So the disablereuse solution you suggest is the only solution that guarantees us we won't redirect clients to the wrong machine (at the expense of a DNS query). Anyway, thanks for your help, really appreciated.
    – Gustau Pérez
    Aug 25, 2017 at 14:43
-1

For your case better to use nginx or haproxy. Apache to heavy for proxying queries.

Nginx can set resolver like:

server {
    ...
    resolver 127.0.0.1;
    set $backend_upstream "http://dynamic.example.com:80";
    proxy_pass $backend_upstream;
    ...
}

You can read about timeouts in haproxy here, and timeouts in nginx here. Hot to setup nginx.

Improve this answer
4
  • You need NGINX Plus to fix the problem found by @Gustau Pérez. See the option resolve at nginx.com/products/on-the-fly-reconfiguration/#dns
    – Mircea Vutcovici
    Aug 24, 2017 at 13:27
  • 2
    Could you please explain why and when Apache is "heavy" for proxying?
    – Mircea Vutcovici
    Aug 24, 2017 at 13:28
  • Because it uses more memory than nginx. Here is an example: help.dreamhost.com/hc/en-us/articles/… Nginx can handle more requests, while spending less time and memory, so it is perfectly used to distribute statics files and proxy requests.
    – Quarind
    Aug 24, 2017 at 14:01
  • 1
    Thank you for your advice, but we need Apache (it's a project requirement). We know nginx but, IIRC, only the plus version does what we want.
    – Gustau Pérez
    Aug 25, 2017 at 14:40

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .