Performance issue -- serving from cache
Brad House <[email protected]> Sat, 14 Feb 2026 14:45:25 -0500
| Newsgroups | gmane.comp.web.squid.general |
|---|---|
| Message-ID | <[email protected]> |
This is a multi-part message in MIME format.
--===============0013126851679512490==
Content-Type: multipart/alternative;
boundary="------------tR2JgFRVew8U5zvqGEvSDmyw"
Content-Language: en-US
This is a multi-part message in MIME format.
--------------tR2JgFRVew8U5zvqGEvSDmyw
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
I've got a squid deployment where serving from cache can be slower than
an uncached download. I'm seeing speeds of around 50MB/s when serving
from cache, which is much slower than anticipated. Infact, when hitting
fast upstream servers, serving of a non-cached asset is faster (even
though its still hitting squid to fetch it).
I'm thinking there's got to be something wrong with my squid
configuration, I'm currently running on Rocky Linux 10 with Squid 6.10-6.
The VM I'm using currently has 4 cores, 16G RAM and 100G of usable
space. I used fio to measure disk performance and I got
* Random Write: 3629MiB/s (1MB block), 33.2k (4k block) IOPS
* Random Read: 8391MiB/s (1MB block), 43.5k (4k block) IOPS
Adding more cpu cores or ram doesn't appear to impact performance.
The underlying infrastructure is made up of hypervisors with dual 100G
uplinks, both the client and squid run on the same hypervisor cloud.
Network performance is not the issue.
As a test, I spun up Apache Traffic Server and get over 800MB/s when
serving from cache.
We have a large on site build system that spins up runners for GitHub
actions, and they're constantly fetching large assets from the internet
for each build, hence our desire for a caching proxy. We'd rather not
switch to Apache Traffic Server as that doesn't have SSL bump capability
(we haven't yet enabled that capability in squid, however). Hopefully
there's a simple configuration I'm missing.
Just for testing I was pulling large image via http that is below my max
object size:
http://mirrors.edge.kernel.org/ubuntu-releases/20.04.6/ubuntu-20.04.6-live-server-amd64.iso
Configuration below:
acl public src 0.0.0.0/0
acl SSL_ports port 443
acl Safe_ports port 80
acl Safe_ports port 443
http_access deny !Safe_ports
http_access deny CONNECT !SSL_ports
http_access allow localhost manager
http_access deny manager
http_access allow public
http_access deny to_localhost
http_access deny to_linklocal
http_access deny all
http_port 8080
maximum_object_size 2 GB
cache_dir aufs /var/spool/squid 325632 16 256
cache_mem 1000 MB
maximum_object_size_in_memory 102400 KB
coredump_dir /var/spool/squid
refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
refresh_pattern deb$ 129600 100% 129600
refresh_pattern udeb$ 129600 100% 129600
refresh_pattern tar.gz$ 129600 100% 129600
refresh_pattern tar.xz$ 129600 100% 129600
refresh_pattern tar.bz2$ 129600 100% 129600
refresh_pattern \/(Packages|Sources)(|\.bz2|\.gz|\.xz)$ 0 0% 0 refresh-ims
refresh_pattern \/Release(|\.gpg)$ 0 0% 0 refresh-ims
refresh_pattern \/InRelease$ 0 0% 0 refresh-ims
refresh_pattern \/(Translation-.*)(|\.bz2|\.gz|\.xz)$ 0 0% 0 refresh-ims
refresh_pattern changelogs.ubuntu.com\/.* 0 1% 1
Thanks!
-Brad
--------------tR2JgFRVew8U5zvqGEvSDmyw
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: 8bit
<!DOCTYPE html>
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body>
<p>I've got a squid deployment where serving from cache can be
slower than an uncached download. I'm seeing speeds of around
50MB/s when serving from cache, which is much slower than
anticipated. Infact, when hitting fast upstream servers, serving
of a non-cached asset is faster (even though its still hitting
squid to fetch it). </p>
<p>I'm thinking there's got to be something wrong with my squid
configuration, I'm currently running on Rocky Linux 10 with Squid
6.10-6.</p>
<p>The VM I'm using currently has 4 cores, 16G RAM and 100G of
usable space. I used fio to measure disk performance and I got </p>
<ul data-list-tree="true" data-indent="1" data-border="0"
style="margin: 4px 0 4px 24px; padding: 0; list-style-position: outside">
<li style="margin: 2px 0; padding: 0; color: #1d1c1d">Random
Write: 3629MiB/s (1MB block), 33.2k (4k block) IOPS</li>
<li style="margin: 2px 0; padding: 0; color: #1d1c1d">Random Read:
8391MiB/s (1MB block), 43.5k (4k block) IOPS</li>
</ul>
<p>Adding more cpu cores or ram doesn't appear to impact
performance.</p>
<p>The underlying infrastructure is made up of hypervisors with dual
100G uplinks, both the client and squid run on the same hypervisor
cloud. Network performance is not the issue.</p>
<p>As a test, I spun up Apache Traffic Server and get over 800MB/s
when serving from cache.</p>
<p>We have a large on site build system that spins up runners for
GitHub actions, and they're constantly fetching large assets from
the internet for each build, hence our desire for a caching
proxy. We'd rather not switch to Apache Traffic Server as that
doesn't have SSL bump capability (we haven't yet enabled that
capability in squid, however). Hopefully there's a simple
configuration I'm missing.</p>
<p>Just for testing I was pulling large image via http that is below
my max object size:
<a class="moz-txt-link-freetext" href="http://mirrors.edge.kernel.org/ubuntu-releases/20.04.6/ubuntu-20.04.6-live-server-amd64.iso">http://mirrors.edge.kernel.org/ubuntu-releases/20.04.6/ubuntu-20.04.6-live-server-amd64.iso</a></p>
<p>Configuration below:</p>
<p>acl public src 0.0.0.0/0<br>
acl SSL_ports port 443<br>
acl Safe_ports port 80<br>
acl Safe_ports port 443<br>
http_access deny !Safe_ports<br>
http_access deny CONNECT !SSL_ports<br>
http_access allow localhost manager<br>
http_access deny manager<br>
http_access allow public<br>
http_access deny to_localhost<br>
http_access deny to_linklocal<br>
http_access deny all<br>
http_port 8080<br>
maximum_object_size 2 GB<br>
cache_dir aufs /var/spool/squid 325632 16 256<br>
cache_mem 1000 MB<br>
maximum_object_size_in_memory 102400 KB<br>
coredump_dir /var/spool/squid<br>
refresh_pattern -i (/cgi-bin/|\?) 0 0% 0<br>
refresh_pattern deb$ 129600 100% 129600<br>
refresh_pattern udeb$ 129600 100% 129600<br>
refresh_pattern tar.gz$ 129600 100% 129600<br>
refresh_pattern tar.xz$ 129600 100% 129600<br>
refresh_pattern tar.bz2$ 129600 100% 129600<br>
refresh_pattern \/(Packages|Sources)(|\.bz2|\.gz|\.xz)$ 0 0% 0
refresh-ims<br>
refresh_pattern \/Release(|\.gpg)$ 0 0% 0 refresh-ims<br>
refresh_pattern \/InRelease$ 0 0% 0 refresh-ims<br>
refresh_pattern \/(Translation-.*)(|\.bz2|\.gz|\.xz)$ 0 0% 0
refresh-ims<br>
refresh_pattern changelogs.ubuntu.com\/.* 0 1% 1</p>
<p><br>
</p>
<p>Thanks!</p>
<p>-Brad</p>
</body>
</html>
--------------tR2JgFRVew8U5zvqGEvSDmyw--
--===============0013126851679512490==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
_______________________________________________
squid-users mailing list
[email protected]
https://lists.squid-cache.org/listinfo/squid-users
--===============0013126851679512490==--