Source Repository:Repository:Gentoo Portage Tree

Summary: Varnish is a state-of-the-art, high-performance HTTP accelerator

Use Flags

Build additional tools: varnishhist, varnishstat, varnishtop, varnishsizes
Use dev-libs/jemalloc for allocations



Project Unfork Status

Here's an update on Project Unfork, plus other neat things.
2015-10-03 by Drobbins

IP Space Migration Continues

All Funtoo user containers in the 8.28 IP space will be moving into our new IP space (172.97) over the next few days. If you have DNS set up -- be sure to watch your container and update to the new IP! DNS will be updated after the move.
2015-08-27 by Drobbins

Funtoo Hosting IP Move

Funtoo user containers with IPs in the 72.18.x.x range will be gradually migrating to new IP addresses this week. If you have DNS entries for your containers, please be aware that your DNS will need to be updated.
2015-08-11 by Drobbins



We welcome improvements to this page. To edit this page, Create a Funtoo account. Then log in and then click here to edit this page. See our editing guidelines to becoming a wiki-editing pro.

Varnish is a Web cache and HTTP accelerator. Varnish will either serve cached content, or retrieve content from the server, and cache it. Varnish cache will reduce I/O pressure on Web servers.



Install Package:Varnish:

# emerge www-servers/varnish



as your varnish is local to your server, your server needs to be aware that it is behind a proxy, and configured for x-forwarded-for or similar to fetch real users IP addresses instead of displaying or localhost.

Configuration is controlled by /etc/varnish/default.vcl & /etc/conf.d/varnishd

/etc/varnish/default.vcl - varnish configuration file
vcl 4.0;
backend default {
    .host = "";
    .port = "8080";
/etc/conf.d/varnishd - varnish configuration file
VARNISHD_OPTS="${VARNISHD_OPTS} -u varnish -g varnish"

Varnish will fetch data from localhost:8080 and serve accelerated proxy data on localhost:80

Achieving a high hit rate


For 10,000 concurrent connections -- or not -- a few configuration settings control the power of varnish. If you're having varnish directly serve to the outside world, dial back concurrency to say 50 or 100 connections per IP. You must take into account corporations and universities hammering several connections from a singular IP. c10k is useful information for if varnish is behind a load balancer such as pound, nginx, or tengine, and all requests are internal and local.

/etc/conf.d/varnishd - varnish concurrency settings
VARNISHD_OPTS="-a -p thread_pool_min=20 -p thread_pool_max=1000 -p thread_pool_add_delay=2 -s malloc,700M"
/etc/varnish/default.vcl - varnish concurrency settings
backend default {
    .host = "";
    .port = "8080";
    .connect_timeout = 600s;
    .first_byte_timeout = 600s;
    .between_bytes_timeout = 600s;
    .max_connections = 10000;
    # .port = "80" led to issues with competing for the port with apache.

Boot Service

To start varnish immediately:

# rc-service varnishd start

To start varnish at boot:

# rc-update add varnishd default


To verify that your traffic is going through varnish, and see if you're getting hits or misses:

$ curl -I


Package:Apache-tools apache benchmark can show the power of varnish. The examples shown are running 500 requests with concurrency of 100 hits.

Example 1
ab against a 3 worker cluster mode puma server
# ab -n 500 -c 100 | grep Request

Requests per second: 110.92 [#/sec] (mean)

Example 2
ab against the same server served through varnish
# ab -n 500 -c 100 | grep Request

Requests per second: 10268.42 [#/sec] (mean)

SSL support

Varnish does not support ssl. There are packages to get around this limitation:

  • net-misc/stunnel (package not on wiki - please add)
  • www-servers/pound (package not on wiki - please add)


External Resources