Squid developers are striving (some may say “with varying levels of success”) to have a clean and safe codebase. One of the available tools is to rely on the compiler and instruct it to treat all warnings as errors to ensure all possibly meaningful signal on issues is caught and acted on. This is done with the -Wall -Werror compiler flags
However, we are not using these flags in ./configure – we set it in SQUID_CXXFLAGS which is a custom variable that’s later refrenced by Makefile.am and used during the build phase. Why is that?
The answer is: doing so would break key autoconf macros such as AC_CHECK_LIB, AC_SEARCH_LIBS, or AC_CHECK_FUNCS. These macros try to compile and link a test executable which calls a function (say, log defined in -lm ) with a standard signature: returning char and having no arguments. Having -Werror turned on while using these functions will result in (tested with gcc 11.4):
configure:52533: gcc -o conftest -Wall -g -O2 -Werror -g conftest.c -lm >&5 conftest.c:366:6: error: conflicting types for built-in function 'log'; expected 'double(double)' [-Werror=builtin-declaration-mismatch] 366 | char log (); | ^~~ conftest.c:1:1: note: 'log' is declared in header '<math.h>' 1 | /* confdefs.h */ cc1: all warnings being treated as errors
CFLAGS="-Werror -Wall -Wno-builtin-declaration-mismatch" ./configure will force ignoring this specific error and ensure that everything builds.
This is GCC specific; Clang doesn’t support this flag and will not exhibit the same behaviour
I could not find any mention of this in a quick internet search; hopefully this writeup will save someone else a bit of time and experimentation
Once an entire dependency tree is exploded, a single c++ include file can become huge, and easily span tens if not hundreds of source files; these will need to be parsed for each compilation unit (c++ file), resulting in a large amount of duplicate work.
So compiler writers came up with the clever idea to optionally save an intermediate state of key headers, to reduce the amount of duplicate work. gcc , clang, msvc all support some variant of the precompiled headers idea.
How do they work in practice?i
Each compiler has its own quirks.
On GCC
A precompiled header has the same name as the header it accelerates with an additional .gch suffix, placed in the same directory as the header file it refers to. It is generated by calling the compiler with the same exact command line arguments as used to build the code, with the additional switches -x c++-header . If a precompiled header is present, it will be automatically used
On clang
A precompiled header has the same name as the header it accelerates with an additional .pch suffix. It is generated by calling the compiler with the same arguments as used to build the code, with the additional switches -x c++-header -emit-pch (the latter might be implicit if the former is supplied). To use it, it is not enough that it be present; the compiler switch -include-pch <pch-file-path> must be used.
On top of this: clang internal documentation highlights that there can only be one precompiled header and it must be included at the beginning of the translation unit
On MSVC
This is not yet a specific target for squid
Could it work for squid?
Yes, in theory. Our coding guidelines mandate that each c++ file start including “squid.h”, which in turn includes our whole portability abstraction layer, which in turn takes in several system files. On my Ubuntu Linux system, a total of 206 header files have to be read and parsed just for this purpose for each of the over 800 files that make up squid. Sounds promising!
Does it work for squid?
In short, unfortunately not. I have experimented with a feature branch, and the results are not what I was hoping for, under several dimensions.
The good: performance gains
I ran some checks, on a NUC6i7KYB (Intel(R) Core(TM) i7-6770HQ CPU @ 2.60GHz, with 16 GiB core, SSD). The test command was
git clean -fdx && ./bootstrap.sh && ./configure && time (make -s -j12 all && make -s -j12 check)
Over 6 attempts, wall clock time averaged 10 minutes and 18 seconds without precompiled headers, and 9 minutes and 48 seconds with them, so a roughly 30 seconds (or about 5%) compile time improvement with gcc. Good but not earth shaking.
The bad: poor integration with the autotools toolchain (gcc edition)
Autotools’ stance on precompiled headers is pretty clear:
This is how I’ve done it. It’s hacky, but some parts of it may not apply to other projects’ setup.
In configure.ac, define an user argument --enable-precompiled-headers , and react to it with an automake conditional ENABLE_PCH_GCC .
In src/Makefile.am , define a custom Makefile rule that builds the precompiled header:
src/Makefile.am is touching files in include/ This is necessary because include/ doesn’t have a Makefile.am of its own, and the top level Makefile.am doesn’t have access to the CXXCOMPILE variable.
srcdir shouldn’t be mucked about at build time; that’s what builddir is for
Then, add a section
if ENABLE_PCH_GCC PCH_FILE=$(top_srcdir)/include/squid.h.gch endif
which is then referenced in
BUILT_SOURCES = \
dnl ... \
$(PCH_FILE)
This will pull in the precompiled header in the list of dependencies of squid, unit tests, and files to clean up. We can’t really control the order this gets built in, but it isn’t a big deal: if we need to compile anything before the precompiled header is built, everything will still work, just without the speed bump.
The worse: poor integration with the autotools toolchain (clang edition)
Clang has one extra problem compared to gcc: to actually use the precompiled header, it needs the -include-pch <file> option. If the option is used, the file needs to be there, or the build will fail.
Which makes being unable to control the build order a showstopper. We would need to build the precompiled header file without that flag before we do anything else. But looking at the generated src/Makefile:
One way to make sure that we only add the –include-pch option would be to send it down the recursive make invocation, except we don’t control that.
That’s it, I give up
The benefit is just not worth the number of hacks and complexity.
What could make it work?
gcc gets this behaviour right; it would be great if clang took inspiration from them. At the very least, do not fail building if the included pch was missing. This would enable treating it for what it is: an optimisation.
My main contribution to the Squid Web Cache project is these days running the project’s infrastructure. A lot of it is the project’s CI/CD farm.
In order to run it, we rely on a very kind donation by DigitalOcean . We use a VM hosted there to run the main jenkins instance and part of the jobs for the x86-64 architecture. We are then using the jenkins digitalocean plugin to spin up instances (droplets) on demand when we need to have more throughput from our build jobs.
In order to maximise how we use our resources, we rely on docker to run all of our target linux userlands. This allows us to decouple the runtime environment from the machine that’s running it, and to ensure consistency across builds (also coming up: a proper staging system).
In this post I’ll focus on how we spin up these instances, the whole setup is a bit more convoluted.
The digitalocean plugin is quite well integrated and easy to use; TBH I haven’t tried plugins for EC2 or GCP, but my other reference point, jclouds, was much harder to configure and set up.
Given our prerequisites, we need ondemand instances to only contain the docker runtime and java, which is needed to run the jenkins slaves as unlike other setups I’ve found online, these run outside the docker containers.
In order to do that, we supply to the “User Data” section this config snippet:
#cloud-config
apt_upgrade: true
package_upgrade: true
packages:
- openjdk-11-jre-headless
- docker.io
users:
- name: name of the jenkins user on the executor machine
groups: docker
shell: /bin/bash
ssh-authorized-keys: ssh-rsa ssh public key of the user jenkins runs under
These actions will be run when the droplet is launched, and prep the executor for jenkins to ssh into it and run the test jobs. In order to give the droplet time to do that, we need to wait for it with this init script:
#!/bin/bash
echo "starting init script"
while ! cloud-init status|grep -qF 'done'
do
echo "waiting for cloud-init to be done"
sleep 10
done
The next tricky bit is in the Droplet section, in the node Labels section we define a label for triggering the instance startup when needed. It can be anything, in our case docker-build-host, and an instance cap.
Referencing this label in the projects’ configuration matrix will trigger the spinup and imaging. Jenkins will then connect to the droplet via ssh and use docker run commands to test the various runtime environments
The Squid Wiki is hosted on an own instance of MoinMoin. We picked it at the time as it had fewer external dependencies than other engines, and it fit the bill.
Over time, and as the number of pages grew, its strengths became limitations, and I’m currently exploring whether to switch to a different engine. Mediawiki is the go-to choice for most people, so that’s what I investigated first.
W3C has developed a tool to convert from one to the other, but it hasn’t been updated in some time, to the point where MediaWiki API changes have bit-rotten it. It doesn’t help that this tends to be a one-off activity, so it
Open Source to the rescue! I have patched it to support current API and it worked for me ™. While waiting for the PR to be approved, feel free to use my fork of it
Transaction Annotations is a feature added in Squid-3.4 which is being used solve some of the more annoying and difficult old problems with Squid configuration and performance. But it seems as yet has not made much of a splash in general usage.
The basic concept of these annotations grows out of the old external ACL helpers tag= feature. Originally the ACL helpers could add one tag to the client request state data and later ACLs could test for its value even in the “fast” type ACL checks without having to repeat any complex work the helper already did or risking unreliable match results.
With Squid-3.4 we took this nice little feature and extended it to the extreme.
the helper protocols got altered so any helper could produce key=value details and send them back to Squid. But not only that; they can send any key name not just ‘tag’, they can send multiple keys and even repeat one key multiple times.
the note directive was added so admin could configure some requests to always be marked with certain annotations.
logging codes were added to record annotations individually, or in groups to the log files.
a note ACL type to match these helper annotations. Replacing the original tag ACL type.
With Squid-4 external ACL have also been updated to accept any logformat code. Many of these are useful in themselves, but for this article we shall focus on the note format code.
Use Case #1: Re-checking authentication without 407 or 401
It is easy to find oneself writing access permissions that require testing the username but not wanting the client to be presented with a 407/401 or popup if the test fails.
In the past the only way to do this was to use a proxy_auth ACL with explicit username(s) listed. Followed by a non-authentication ACL test or the ‘all hack’.
With annotations there is now a third option. After an initial authentication ACL check has been passed the user= annotation has been added to the request. Simply checking the note ACL to test for the user key having been supplied by an authentication (or external ACL) helper with value being whatever username(s) to match.
Use Case #2: User based outgoing IP (or TOS)
Making Squid use a particular outgoing IP or TOS for one user but not for others has been very difficult for almost the whole exitence of Squid. If each user had an identifiable source IP it was not too bad, but once they used a downstream proxy all bets were off. The tcp_outgoing_* directives simply do not support helper lookups reliably.
With annotations, we can use the note ACL in a tcp_outgoing_addr or tcp_outgoing_tos access list to test for the user key having been supplied by an authentication (or external ACL) helper with value being whatever username we want to match.
Use Case #2: Fast group ACLs
In complex networks with many user groups being assigned and controlling different Squid functionality one may find oneself trying to optimize performance for a large number of separate external ACLs which only check for the users membership of a group.
Remembering that these are slow async lookups, and the resulting restriction to asynchronous (slow group) access controls can lead to administrative problems and some annoying workarounds in squid.conf.
With annotations, we can reduce the group lookup to a single helper query which returns a list of group=X annotations. Then use the note ACL again to test without any fast/slow group access control restrcictions.
If you are using a custom authenticator you could then even have it produce that list of groups alongside the user= credentials. Completely avoiding the need for an external ACL helper.
We are currently weathering both a PSU meltdown and disk failure (with full compliment of side effects) on the main squid-cache.org server. The Squid sysadmin and Foundation are all working on it as fast as possible.
Donations towards the purchase of a new server are greatly appreciated and will allow us to buy better hardware.
UPDATE: 2014-10-08: after weeks of late nights and very early mornings by the volunteer sysadmin team we are nearly all back up to full operational state again. The latest status of each affected major service is outlined below.
Mail and Mailing lists
The mail server for squid-cache.org was worst hit. Email has been down across most of September since the initial event. Any email sent to a squid-cache.org email address will have been held up and/or bounced.
Mail services are now back, but some spam control upgrades were forced on us that are still having fallout. Mailing lists are being migrated to a new domain name lists.squid-cache.org. Existing subscriptions have been automatically moved to the new list domain. You can expect to see an initial post explaining the change when the list you are subscribed to is recovered.
This change may require updates to mail filtering and rules outside our services. If you are aware of any in your domain or systems please see that they are updated.
IMPORTANT: some subscriptions have subsequently had to be removed due to backscatter spam from relays and corporate ticket logging systems. Posters to the list know who I am talking about. If you find your subscription has gone silent again recently please check the systems you are having mail delivered to and through then try re-subscribing to the new list.
Mail archives are currently split between the old hypermail + swift archival system and new pipermail. This is reflected on the website. If you are maintaining a mirror of the Squid mail archives please subscribe to our new mailing list for mirror operators and get in touch with the sysadmin team to sort out what is going to happen with mail mirrors in future.
DNS
We believe this is recovered. If anyone is still having issues resolving the domains please get in touch with noc @ lists.squid-cache.org.
Code Repository
The repository has been fully recovered and service on bzr.squid-cache.org and rsync is resumed.
FTP
The FTP service has been limping along with access but no updates. The main server is now in the process of being rebuilt from scratch. Please do not be surprised if you are suddenly challenged for login, try a mirror instead. Anonymous access to the main FTP will be resumed ASAP.
Website
The http://www.squid-cache.org site is mostly up and running. Mirrors have remained available for the duration, but were not being updated with daily contents. The updates should now have resumed, but there are still a few kinks to work out in the content. If you find any issues going forward please report it in our bugzilla under Project Services.
Mirror Services and rsync
If you are running a WWW mirror please ensure you are using rsync access and your server is capable of serving the http://www.squid-cache.org name as outlined in the mirror guidelines. Similar goes for FTP mirrors. We are adding a new mailing list for mirror server contacts. Our database of registered contacts for HTTP and FTP mirrors will be automatically subscribed so please keep an eye on the mailbox you registered with us already. Anyone running a Squid mirror of any kind please subscribe and post your mirror details to the list.
The rsync service itself is running with some data shares temporarily disabled. These will be re-opened as the services are brought back to full functionality. There are no changes to remote configurations provided you have been following the current mirror guidelines. The dynamic website (http-files-dyn) will no longer be publicly available, please mirror the static (http-files) instead.
Apologies for the inconvenience.
… and Murphys Law has not finished with us yet:
Some security vulnerabilities were reported. A new squid-3.4.8 package has been released to resolve those. All users relying on SNMP or the pinger helper are advised to upgrade. The SNMP details can be found here, pinger details can be found here.
If you have not been aware of the IETF HTTPbis Working Group and what we do, it is chartered to improve HTTP. For the last decade and a half HTTP/1.1 has been defined by the monolithic and sometimes confusing RFC2616 document with a relatively few extensions. The WG has been putting in a lot of effort to simplify the texts and clarify how the protocol actually works.
If you have been putting off reading the HTTP/1.1 specification because of its enormous length now is a good time to dive in. The text has never been simper and easier to read. Changes from the old document have been kept minimal, but there are some listed in the Appendices.
Mark Nottingham the WG chairman made this formal announcement a few hours ago:
The revision of HTTP/1.1’s specification, obsoleting RFC2616, is complete.
Thanks to everyone who has commented upon, reviewed and otherwise contributed to them over this nearly seven-year(!) effort.
Special thanks to our Area Directors over the years: Lisa Dusseault, Alexey Melnikov, Peter Saint-Andre and Barry Leiba, along with Yves Lafon, who helped edit Range Requests.
Finally, please warmly thank both Roy Fielding and Julian Reschke the next time you see them (I believe beer would be appreciated); the amount of effort that they put into these documents is far, far more than they originally signed up for, and they’ve done an excellent job.
The Squid Software Foundation is seeking to expand the board of directors. We currently have three directors and looking for at least one more to join the team. For details about the position and what the directors do please see http://www.squid-cache.org/Foundation/director.html
Being a Squid Software Foundation Director is a serious responsibility, but also a cool gig! Not only can you have an immediate and significant impact on the Squid Project, but you can earn admiration and respect of your peers while doing more than just your usual software development, system administration, or support activities.
Do you want to brag about being more than a successful geek? Exercise the parts of your brain you did not know you had? Resolve real-world conflicts and balance real-world trade-offs? Then how about solving a few difficult Squid Project problems? Want to spice up your resume or simply learn to manage a popular open source project? Consider nominating yourself!
Applicants should contact board@squid-cache.org with nomination for the position of Director. Self-nominations are accepted and encouraged. Please indicate why you think the nominee would be a good Foundation director.
Please submit nominations by October 4th, 2013.
The Squid Software Foundation Board of Directors
Henrik Nordström,
Amos Jeffries,
Alex Rousskov.
One of the new features brought in with Squid-3.2 is dynamic helpers. A brief name for a very useful administrative tool and like all tools can be both easy and tricky to use at the same time.
If you have a proxy using helper processes but only a small cache (or none) this is a feature for you.
The good news
Configuration is super easy – just set the initial startup, maximum number of helpers and an idle value for how many to start when new ones are needed to handle the request load.
Dying helpers have a higher threshold before they kill Squid. It is not perfectly tuned yet in the code so improvements will contnue to happen here, but already we see small bursts of helper failures being suppressed by re-started replacements without that all too familiar Squid halt with “helper dying too quickly” messages. Note that this is just higher, not gone.
The bad news
Determining what those values should be is no more easy or straightforward than before. Squid uses fork() to start new helpers. The main side effect of this is that helper instances started while Squid is running will require a virtual memory size equivalent to the Squid worker process memory at the time they are started. If your proxy is pushing the box to its limit on RAM, dynamically started helpers could easily push it over to swapping memory at the worst possible time (peak load starting to arrive). Also on the bad news side is that the helpers are run per-worker. Which has the potential to compound the RAM usage problems.
We do have a proposal put to the development team which will almost completely remove this problem. Having the coordinator or a special spawner kid do the forking instead of the heavy workers. But as of this writing nobody is working on it (volunteers welcome, please contact the squid-dev mailing list).
Practice Guidelines
While it may look like the bad news is worse than the good news it does turn out that most installations are small instances or non-caching worker proxies these days. All of which may need lots of helpers, but are not heavy on the RAM requirements. For all these installations dynamic helpers are ideal and in a lot of cases can even be set with zero helpers on startup for a very speedy delay to first request accepted time.
The caching proxy installations with higher memory requirements in the workers can still make use of the dynamic nature to avoid complete outages in the worst-case situations where peak traffic overloads the planned helpers. But should normally be configured as before with enough helpers to meet most needs started before the RAM requirements become too onerous on the worker.
Until at least the bad news problems above are resolved the default behaviour for Squid will continue to be starting all the maximum helpers on startup. So there are no unexpected surprises for upgrading, and the old advice on calculating helper requirements is still useful for determining that maximum.
The new features for Squid-3.2 are now decided and present, the latest builds seem to be running okay. Operating system distributors are starting to work on producing packages for the upcoming release.
So when do we get to see a stable release?
Yes, well. There is just one little problem. Before 3.2 can be released as stable for widespread production use we need to be sure that there are no serious bugs in the new or updated code. Right now we are aware of a few that have not been fixed.
We need assistance fixing bugs in the 3.2 beta.
The serious bug clearing focus actually began two months ago. The worst bugs have now been squashed and we are down to the last few dozen major bugs blocking a stable release. You can find these marked as major, critical, or blocker in our bugzilla. Any assistance finding the causes or working patches for the remaining bugs is very welcome and will help speed up the release process.
IMPORTANT: please ensure that the bugzilla gets your reports.
What is the fuss about Squid-2.7?
Squid 3.2 is a little bit unusual. Being the release where the Squid-3 series finally superceeds the Squid 2.6 and 2.7 fork in both common features and performance. Squid-2 has not been actively maintained for more than a year now. Features available in that alternate series of Squid are almost all available in Squid-3.2, the remaining features are expected to be ported over shortly after 3.2 is released stable and developer time becomes more available.
What this means in terms of bugs is that a lot of the 2.6 and 2.7 series bugs are being closed with target milestone of 3.2 when they are fixed or no longer relevant to 3.2 code. So if you are waiting for a 2.7 series bug to be closed, please do not be alarmed when its closed against 3.2 without a 2.7 fix being available.
We expect 3.2 to be useful wherever Squid 2.7 is currently running, if you find the upgrade not working that is a problem we need to resolve as soon as possible. So please give it a try and report problems. Just remember to read the 3.2 release notes carefully, and possibly the 3.1 release notes as well.
By and large these older squid-2 series bugs are not going to block the 3.2 release any more than old 3.0 and 3.1 bugs will. But identifying and closing bugs no longer relevant will benefit everyone by allowing us to focus more on the bugs which are still biting people.
There are also hundreds of minor bugs which can be worked on as well.
The Squid Web Proxy/Cache Blog 
You must be logged in to post a comment.