Posts Tagged ‘sysadmin’

Keeping Your NodeJS Web App Running on Production Linux

June 27, 2015

All the following offerings that I’ve evaluated target different scenarios. I’ve listed the pros and cons for each of them and where I think they fit into a potential solution to monitor your web applications (I’m leaning toward NodeJS) and make sure they keep running. I’ve listed the goals I was looking to satisfy.

For me I have to have a good knowledge of the landscape before I commit to a decision and stand behind it. I like to know I’ve made the best decision based on all the facts that are publicly available. Therefore, as always, it’s my responsibility to make sure I’ve done my research in order to make an informed and ideally… best decision possible. I’m pretty sure my evaluation was un-biased, as I hadn’t used any of the offerings other than forever before.

I looked at quite a few more than what I’ve detailed below, but the following candidates I felt were worth spending some time on.

Keep in mind, that everyone’s requirements will be different, so rather than tell you which to use because I don’t know your situation, I’ve listed the attributes (positive, negative and neutral) that I think are worth considering when making this choice. After my evaluation I make some decisions and start the configuration.

Evaluation criterion

  1. Who is the creator. I favour teams rather than individuals, as individuals move on, then where does that leave the product?
  2. Does it do what we need it to do? Goals address this.
  3. Do I foresee any integration problems with other required components?
  4. Cost in money. Is it free? I usually gravitate toward free software. It’s usually an easier sell to clients and management. Are there catches once you get further down the road? Usually open source projects are marketed as is.
  5. Cost in time. Is the set-up painful?
  6. How well does it appear to be supported? What do the users say?
  7. Documentation. Is there any / much? What is it’s quality?
  8. Community. Does it have an active one? Are the users getting their questions answered satisfactorily? Why are the unhappy users unhappy (do they have a valid reason).
  9. Release schedule. How often are releases being made? When was the last release?
  10. Gut feeling, Intuition. How does it feel. If you have experience in making these sorts of choices, lean on it. Believe it or not, this should probably be No. 1.

The following tools have been my choice based on the above criterion.

Goals

  1. Application should start automatically on system boot
  2. Application should be re-started if it dies or becomes un-responsive
  3. Ability to add the following later without having to swap the chosen offering:
    1. Reverse proxy (Nginx, node-http-proxy, Tinyproxy, Squid, Varnish, etc)
    2. Clustering and providing load balancing for your single threaded application
    3. Visibility of application statistics.
  4. Enough documentation to feel comfortable consuming the offering
  5. The offering should be production ready. This means: mature with a security conscious architecture.

Sysvinit, Upstart, systemd & Runit

You’ll have one of these running on your Linux box.

These are system and service managers for Linux. Upstart and the later systemd were developed as replacements for the traditional init daemon (Sysvinit), which all depend on init. Init is an essential package that pulls in the default init system. In Debian, starting with Jessie, systemd is your default system and service manager.

There’s some quite helpful info on the differences between Sysvinit and systemd here.

systemd

As I have systemd installed out of the box on my test machine (Debian Jessie), I’ll be using this for my set-up.

Documentation

  1. Well written comparison with Upstart, systemd, Runit and even Supervisor.

Running the likes of the below commands will provide some good details on how these packages interact with each other:

aptitude show sysvinit
aptitude show systemd
# and any others you think of

These system and service managers all run as PID 1 and start the rest of the system. Your Linux system will more than likely be using one of these to start tasks and services during boot, stop them during shutdown and supervise them while the system is running. Ideally you’re going to want to use something higher level to look after your NodeJS app. See the following candidates…

forever

and it’s web UI. Can run any kind of script continuously (whether it is written in node.js or not). This wasn’t always the case though. It was originally targeted toward keeping NodeJS applications running.

Requires NPM to install globally. We already have a package manager on Debian and all other main-stream Linux distros. Installing NPM just adds more attack surface area. Unless it’s essential, I’d rather do without NPM on a production server where we’re actively working to reduce the installed package count and disable everything else we can. I could install forever on a development box and then copy to the production server, but it starts to turn the simplicity of a node module into something not as simple, which then makes offerings like Supervisor, Monit and Passenger look even more attractive.

NPM Details

Does it Meet our Goals?

  1. Not without an extra script. Crontab or similar
  2. Application will be re-started if it dies, but if it’s response times go up, there’s not much forever is going to do about it. It has no way of knowing.
  3. Ability to add the following later without having to swap the chosen offering:
    1. Reverse proxy: I don’t see a problem
    2. Integrate NodeJS’s core module cluster into your NodeJS application for load balancing
    3. Visibility of application statistics could be added later with the likes of Monit or something else, but if you used Monit, then there wouldn’t really be a need for forever as Monit does the little that forever does and is capable of so much more, but is not pushy on what to do and how to do it. All the behaviour is defined with quite a nice syntax in a config file or as many as you like.
  4. I think there is enough documentation to feel comfortable consuming it, as forever doesn’t do a lot, which doesn’t have to be a bad thing.
  5. The code it self is probably production ready, but I’ve heard quite a bit about stability issues. You’re also expected to have NPM installed (more attack surface) when we already have native package managers on the server(s).

Overall Thoughts

For me, I’m looking for a tool set that does a bit more. Forever doesn’t satisfy my requirements. There’s often a balancing act between not doing enough and doing too much.

PM2

PM2

Younger than forever, but seems to have quite a few more features and does actually look quite good. I’m not sure about production ready though?

As mentioned on the github page: “PM2 is a production process manager for Node.js applications with a built-in load balancer“. This “Sounds” and at the initial glance looks shiny. Very quickly you should realise there are a few security issues you need to be aware of though.

The word “production” is used but it requries NPM to install globally. We already have a package manager on Debian and all other main-stream Linux distros. Installing NPM just adds more attack surface area. Unless it’s essential and it shouldn’t be, I’d rather do without it on a production system. I could install PM2 on a development box and then copy to the production server, but it starts to turn the simplicity of a node module into something not as simple, which then makes offerings like Supervisor, Monit and Passenger look even more attractive.

At the time of writing this, it’s less than a year old and in nodejs land, that means it’s very much in the immature realm. Do you really want to use something that young on a production server? I’d personally advise against it.

Yes, it’s very popular currently. That doesn’t tell me it’s ready for production though. It tells me the marketing is working.

Is your production server ready for PM2? That phrase alone tells me the mind-set behind the project. I’d much sooner see it the other way around. Is PM2 ready for my production server? You’re going to need a staging server for this, unless you honestly want development tools installed on your production server (git, build-essential, NVM and an unstable version of node 0.11.14 (at time of writing)) and run test scripts on your production server? Not for me or my clients thanks.

If you’ve considered the above concerns and can justify adding the additional attack surface area, check out the features if you haven’t already.

Features that Stood Out

They’re also listed on the github repository. Just beware of some of the caveats. Like for the load balancing: “we recommend the use of node#0.11.15+ or io.js#1.0.2+. We do not support node#0.10.* cluster module anymore!” 0.11.15 is unstable, but hang-on, I thought PM2 was a “production” process manager? OK, so were happy to mix unstable in with something we label as production?

On top of NodeJS, PM2 will run the following scripts: bash, python, ruby, coffee, php, perl.

Start-up Script Generation

Although I’ve heard a few stories that this is fairly un-reliable at the time of writing this. Which doesn’t surprise me, as the project is very young.

Documentation

  1. Advanced Readme

Does it Meet our Goals?

  1. The feature exists, unsure of how reliable it is currently though?
  2. Application should be re-started if it dies shouldn’t be a problem. PM2 can also restart your application if it reaches a certain memory threshold. I haven’t seen anything around restarting based on response times or other application health issues.
  3. Ability to add the following later without having to swap the chosen offering:
    1. Reverse proxy: I don’t see a problem
    2. Clustering and load-balancing is integrated but immature.
    3. PM2 provides a small collection of viewable statistics. Personally I’d want more, but I don’t see any reason why you’d have to swap PM2 because of this.
  4. There is reasonable official documentation for the age of the project. The community supplied documentation will need to catch up a bit, although there is a bit of that too. After working through all of the offerings and edge-cases, I feel as I usually do with NodeJS projects. The documentation doesn’t cover all the edge-cases and the development itself misses edge cases. Hopefully with time it’ll get better though as the project does look promising.
  5. I haven’t seen much that would make me think PM2 is production ready. It’s not yet mature. I don’t agree with it’s architecture.

Overall Thoughts

For me, the architecture doesn’t seem to be heading in the right direction to be used on a production web server where less is better. I’d like to see this change. If it did, I think it could be a serious contender for this space.

 


The following are better suited to monitoring and managing your applications. Other than Passenger, they should all be in your repository, which means trivial installs and configurations.

Supervisor

Supervisord

Supervisor is a process manager with a lot of features and a higher level of abstraction than the likes of the above Sysvinit, upstart, systemd, Runit, etc so it still needs to be run by an init daemon in itself.

From the docs: “It shares some of the same goals of programs like launchd, daemontools, and runit. Unlike some of these programs, it is not meant to be run as a substitute for init as “process id 1”. Instead it is meant to be used to control processes related to a project or a customer, and is meant to start like any other program at boot time.” Supervisor monitors the state of processes. Where as a tool like Monit can perform so many more types of tests and take what ever actions you define.

It’s in the Debian repositories  (trivial install on Debian and derivatives).

Documentation

  1. Main web site
  2. There’s a good short comparison here.

Source

Does it Meet our Goals?

  1. Application should start automatically on system boot: Yip. That’s what Supervisor does well.
  2. Application will be re-started if it dies, or becomes un-responsive. It’s often difficult to get accurate up/down status on processes on UNIX. Pidfiles often lie. Supervisord starts processes as subprocesses, so it always knows the true up/down status of its children.If your application becomes unresponsive or can’t connect to it’s database or any other service/resource it needs to work as expected. To be able to monitor these events and respond accordingly your application can expose a health-check interface, like GET /healthcheck. If everything goes well it should return HTTP 200, if not then HTTP 5**In some cases the restart of the process will solve this issue. httpok is a Supervisor event listener which makes GET requests to the configured URL. If the check fails or times out, httpok will restart the process.To enable httpok the following lines have to be placed in supervisord.conf:
  3. Ability to add the following later without having to swap the chosen offering:
    1. Reverse proxy: I don’t see a problem
    2. Integrate NodeJS’s core module cluster into your NodeJS application for load balancing. This would be completely separate to supervisor.
    3. Visibility of application statistics could be added later with the likes of Monit or something else. For me, Supervisor doesn’t do enough. Monit does. Plus if you need what Monit offers, then you have to have three packages to think about, or Something like Supervisor, which is not an init system, so it kind of sits in the middle of the ultimate stack. So my way of thinking is, use the init system you already have to do the low level lifting and then something small to take care of everything else on your server that the init system is not really designed for and Monit has done this job really well. Just keep in mind also. This is not based on any bias. I hadn’t used Monit before this exercise.
  4. Supervisor is a mature product. It’s been around since 2004 and is still actively developed. The official and community provided docs are good.
  5. Yes it’s production ready. It’s proven itself.

 

Overall Thoughts

The documentation is quite good, easy to read and understand. I felt that the config was quite intuitive also. I already had systemd installed out of the box and didn’t see much point in installing Supervisor as systemd appeared to do everything Supervisor could do, plus systemd is an init system (it sits at the bottom of the stack). In most scenarios you’re going to have a Sysvinit or replacement of (that runs with a PID of 1), so in many cases Supervisor although it’s quite nice is kind of redundant, and of course Ubuntu has Upstart.

Supervisor is better suited to running multiple scripts with the same runtime, for example a bunch of different client applications running on Node. This can be done with systemd and the others, but Supervisor is a better fit for this sort of thing.

 

Monit

monit

Is a utility for monitoring and managing daemons or similar programs. It’s mature, actively maintained, free, open source and licensed with GNU AGPL.

It’s in the debian repositories (trivial install on Debian and derivatives). The home page told me the binary was just under 500kB. The install however produced a different number:

After this operation, 765 kB of additional disk space will be used.

Monit provides an impressive feature set for such a small package.

Monit provides far more visibility into the state of your application and control than any of the offerings mentioned above. It’s also generic. It’ll manage and/or monitor anything you throw at it. It has the right level of abstraction. Often when you start working with a product you find it’s limitations and they stop you moving forward and you end up settling for imperfection or you swap the offering for something else providing you haven’t already invested to much effort into it. For me Monit hit the sweet spot and never seems to stop you in your tracks. There always seems to be an easy to relatively easy way to get any monitoring->take action sort of task done. What I also really like is that moving away from Monit should be relatively painless also. The time investment is small and some of it will be transferable in many cases. It’s just config from the control file.

Features that Stood Out

  • Ability to monitor files, directories, disks, processes, the system and other hosts.
  • Can perform emergency logrotates if a log file suddenly grows too large too fast
  • File Checksum TestingThis is good so long as the compromised server hasn’t also had the tool your using to perform your verification (md5sum or sha1sum) modified, which would be common. That’s why in cases like this, tools such as stealth can be a good choice.
  • Testing of other attributes like ownership and access permissions. These are good, but again can easily be modified.
  • Monitoring directories using time-stamp. Good idea, but don’t rely solely on this. time-stamps are easily modified with touch -r … providing you do it between Monit’s cycles and you don’t necessarily know when they are unless you have permissions to look at Monit’s control file.
  • Monitoring space of file-systems
  • Has a built-in lightweight HTTP(S) interface you can use to browse the Monit server and check the status of all monitored services. From the web-interface you can start, stop and restart processes and disable or enable monitoring of services. Monit provides fine grained control over who/what can access the web interface or whether it’s even active or not. Again an excellent feature that you can choose to use or not even have the extra attack surface.
  • There’s also an agregator (m/monit) that allows sys-admins to monitor and manage many hosts at a time. Also works well on mobile devices and is available at a one off cost (reasonable price) to monitor all hosts.
  • Once you install Monit you have to actively enable the http daemon in the monitrc in order to run the Monit cli and/or access the Monit http web UI. At first I thought “is this broken?” I couldn’t even run monit status (it’s a Monit command). ps told me Monit was running. Then I realised… it’s secure by default. You have to actually think about it in order to expose anything. It was this that confirmed Monit for me.
  • The Control File
  • Just like SSH, to protect the security of your control file and passwords the control file must have read-write permissions no more than 0700 (u=xrw,g=,o=); Monit will complain and exit otherwise.

Documentation

The following was the documentation I used in the same order and I found that the most helpful.

  1. Main web site
  2. Official Documentation
  3. Source and links to other documentation including a QUICK START guide of about 6 lines.
  4. Adding Monit to systemd
  5. Release notes

Does it Meet our Goals?

  1. Application can start automatically on system boot
  2. Monit has a plethora of different types of tests it can perform and then follow up with actions based on the outcomes. Http is but one of them.
  3. Ability to add the following later without having to swap the chosen offering:
    1. Reverse proxy: Yes, I don’t see any issues here
    2. Integrate NodeJS’s core module cluster into your NodeJS application for load balancing. Monit will still monitor, restart and do what ever else you tell it to do.
    3. Monit provides application statistics to look at if that’s what you want, but it also goes further and provides directives for you to declare behaviour based on conditions that Monit checks for.
  4. Plenty of official and community supplied documentation
  5. Yes it’s production ready. It’s proven itself. Some extra education around some of the points I raised above with some of the security features would be good. If you could trust the hosts hashing programme (and other commonly trojanised programmes like find, ls, etc) that Monit uses, perhaps because you were monitoring it from a stealth controller (which had already taken a known good copy and produced it’s own bench-mark hash) or similar then yes, you could use that feature of Monit with greater assurance that the results it was producing were in fact accurate. In saying that, you don’t have to use the feature, but it’s there if you want it, which I see as very positive so long as you understand what could go wrong and where.

 

Overall Thoughts

The accepted answer here is a pretty good mix and approach to using the right tools for each job. Monit has a lot of capabilities, none of which you must use, so it doesn’t get in your way, as many opinionated tools do and like to dictate how you do things and what you must use in order to do them. Monit allows you to leverage what ever you already have in your stack. You don’t have to install package managers or increase your attack surface other than [apt-get|aptitude] install monit It’s easy to configure and has lots of good documentation.

Passenger

Passenger

I’ve looked at Passenger before and it looked quite good then. It still does, with one main caveat. It’s trying to do to much. One can easily get lost in the official documentation (example of the Monit install (handfull of commands to cover all Linux distros one page) vs Passenger install (aprx 10 pages)).  “Passenger is a web server and application server, designed to be fast, robust and lightweight. It runs your web apps with the least amount of hassle by taking care of almost all administrative heavy lifting for you.” I’d like to see the actual weight rather than just a relative term “lightweight”. To me it doesn’t look light weight. The feeling I got when evaluating Passenger was similar to the feeling produced with my Ossec evaluation.

The learning curve is quite a bit steeper than all the previous offerings. Passenger has strong opinions that once you buy into could make it hard to use the tools you may want to swap in and out. I’m not seeing the UNIX Philosophy here.

If you look at the Phusion Passenger Philosophy we see some note-worthy comments. “We believe no good software has bad documentation“. If your software is 100% intuitive, the need for documentation should be minimal. Few software products are 100% intuitive, because we only have so much time to develop it. The comment around “the Unix way” is interesting also. At this stage I’m not sure they’ve done better. I’d like to spend some time with someone or some team that has Passenger in production in a diverse environment and see how things are working out.

Passenger isn’t in the Debian repositories, so you would need to add the apt repository.

Passenger is six years old at the time of writing this, but the NodeJS support is only just over a year old.

Features that Stood Didn’t really Stand Out

Sadly there weren’t many that stood out for me.

  • Handle more traffic looked similar to Monit resource testing but without the detail. If there’s something Monit can’t do well, it’ll say “Hay, use this other tool and I’ll help you configure it to suite the way you want to work. If you don’t like it, swap it out for something else” With Passenger it seems to integrate into everything rather than providing tools to communicate loosely. Essentially locking you into a way of doing something that hopefully you like. It also talks about “Uses all available CPU cores“. If you’re using Monit you can use the NodeJS cluster module to take care of that. Again leaving the best tool for the job to do what it does best.
  • Reduce maintenance
    • Keep your app running, even when it crashesPhusion Passenger supervises your application processes, restarting them when necessary. That way, your application will keep running, ensuring that your website stays up. Because this is automatic and builtin, you do not have to setup separate supervision systems like Monit, saving you time and effort.” but this is what Monit excels at and it’s a much easier set-up than Passenger. This sort of marketing doesn’t sit right with me.
    • Host multiple apps at once. Host multiple apps on a single server with minimal effort. ” If we’re talking NodeJS web apps, then they are their own server. They host themselves. In this case it looks like Passenger is trying to solve a problem that doesn’t exist?
  • Improve security
    • Privilege separationIf you host multiple apps on the same system, then you can easily run each app as a different Unix user, thereby separating privileges.“. The Monit documentation says this: “If Monit is run as the super user, you can optionally run the program as a different user and/or group.” and goes on to provide examples how it’s done. So again I don’t see anything new here. Other than the “Slow client protections” which has side affects, that’s it for security considerations with Passenger. From what I’ve seen Monit has more in the way of security related features.
  • What I saw happening here was a lot of stuff that I actually didn’t need. Your mileage may vary.

Offerings

Phusion Passenger is a commercial product that has enterprise, custom and open source (which is free and still has loads of features).

Documentation

The following was the documentation I used in the same order and I found that the most helpful.

  1. NodeJS tutorial (This got me started with how it could work with NodeJS)
  2. Main web site
  3. Documentation and support portal
  4. Design and Architecture
  5. User Guide Index
  6. Nginx specific User Guide
  7. Standalone User Guide
  8. Twitter, blog
  9. IRC: #passenger at irc.freenode.net. I was on there for several days. There was very little activity.

Source

Does it Meet our Goals?

  1. Application should start automatically on system boot. There is no doubt that Passenger goes way beyond this aim.
  2. Application should be re-started if it dies or becomes un-responsive. There is no doubt that Passenger goes way beyond this aim.
  3. Ability to add the following later without having to swap the chosen offering:
    1. Reverse proxy: Passenger provides Integrations into Nginx, Apache and stand-alone (provide your own proxy)
    2. Passenger scales up NodeJS processes and automatically load balances between them
    3. Passenger is advertised as offering easily viewable statistics.
  4. There is loads of official documentation. Not as much community contributed though, as it’s still young.
  5. From what I’ve seen so far, I’d say Passenger is production ready. I would like to see more around how security was baked into the architecture though before I committed to using it.

Overall Thoughts

I spent quite a while reading the documentation. I just think it’s doing to much. I prefer to have stronger single focused tools that do one job, do it well and play nicely with all the other kids in the sand pit. You pick the tool up and it’s just intuitive how to use it and you end up reading docs to confirm how you think it should work. For me, this is not how passenger is.

If you’re looking for something even more comprehensive, check out Zabbix. If you like to pay for your tools, check out Nagios if you haven’t already.


At this point it was fairly clear as to which components I’d be using and configuring to keep my NodeJS application monitored, alive and healthy along with any other scripts or processes. systemd and Monit. If you’re on Ubuntu, you’d probably use Upstart instead of systemd as it should already be your default init system. So going with the default for the init system should give you a quick start and provide plenty of power. Plus it’s well supported, reliable, feature rich and you can manage anything/everything you want without installing extra packages. For the next level up, I’d choose Monit. I’ve now used it in production and it’s taken care of everything above the init system. I feel it has a good level of abstraction, plenty of features, doesn’t get in the way and integrates nicely into your production OS.

Getting Started with Monit

So we’ve installed Monit with an apt-get install monit and we’re ready to start configuring it.

ps aux | grep -i monit

Will reveal that Monit is running:

/usr/bin/monit -c /etc/monit/monitrc

Now if you issue a sudo service monit restart, it won’t work as you can’t access the Monit CLI due to the httpd not running.

The first thing we need to do is make some changes to the control file (/etc/monit/monitrc in Debian). The control file has sensible defaults already. At this stage I don’t need a web UI accessible via localhost or any other hosts, but it still needs to be turned on and accessible by at least localhost. Here’s why:

Note that HTTP support is required for almost all Monit CLI interface operation, as CLI commands (such as “monit status”) are handled by communicating with the Monit background process via the the HTTP interface. So basically you should have this enable, though you can bind the HTTP interface to localhost only so Monit is not accessible from the outside.

In order to turn on the httpd, all you need in your control file for that is:

set httpd port 2812 and use address localhost # only accept connection from localhost
allow localhost # allow localhost to connect to the server and

If you want to receive alerts via email, then you’ll need to configure that. Then on reload you should get start and stop events (when you quit).

sudo monit reload

Now if you issue a curl localhost:2812 you should get the web UI’s response of a html page. Now you can start to play with the Monit CLI

Now to stop the Monit background process use:

monit quit

Oh, you can find all the arguments you can throw at Monit here, or just issue a:

monit -h # will list all options.

To check the control file for syntax errors:

sudo monit -t

Also keep an eye on your log file which is specified in the control file: set logfile /var/log/monit.log

Right. So what happens when Monit dies…

Keep Monit Alive

Now you’re going to want to make sure your monitoring tool that can be configured to take all sorts of actions never just stops running, leaving you flying blind. No noise from your servers means all good right? Not necessarily. Your monitoring tool just has to keep running. So lets make sure of that now.

When Monit is apt-get install‘ed on Debian it gets installed and configured to run as a daemon. This is defined in Monit’s init script.
Monit’s init script is copied to /etc/init.d/ and the run levels set-up for it. This means when ever a run level is entered the init script will be run taking either the single argument of stop (example: /etc/rc0.d/K01monit), or start (example: /etc/rc2.d/S17monit). Further details on run levels here.

systemd to the rescue

Monit is pretty stable, but if for some reason it dies, then it won’t be automatically restarted again.
This is where systemd comes in. systemd is installed out of the box on Debian Jessie on-wards. Ubuntu uses Upstart which is similar. Both SysV init and systemd can act as drop-in replacements for each other or even work along side of each other, which is the case in Debian Jessie. If you add a unit file which describes the properties of the process that you want to run, then issue some magic commands, the systemd unit file will take precedence over the init script (/etc/init.d/monit)

Before we get started, lets get some terminology established. The two concepts in systemd we need to know about are unit and target.

  1. A unit is a configuration file that describes the properties of the process that you’d like to run. There are many examples of these I can show you and I’ll point you in the direction soon. They should have a [Unit] directive at a minimum. The syntax of the unit files and the target files were derived from Microsoft Windows .ini files. Now I think the idea is that if you want to have a [Service] directive within your unit file, then you would append .service to the end of your unit file name.
  2. A target is a grouping mechanism that allows systemd to start up groups of processes at the same time. This happens at every boot as processes are started at different run levels.

Now in Debian there are two places that systemd looks for unit files… In order from lowest to highest precedence, they are as follows:

  1. /lib/systemd/system/ (prefix with /usr dir for archlinux) unit files provided by installed packages. Have a look in here for many existing examples of unit files.
  2. /etc/systemd/system/ unit files created by the system administrator

As mentioned above, systemd should be the first process started on your Linux server. systemd reads the different targets and runs the scripts within the specific target’s “target.wants” directory (which just contains a collection of symbolic links to the unit files). For example the target file we’ll be working with is the multi-user.target file (actually we don’t touch it, systemctl does that for us (as per the magic commands mentioned above)). Just as systemd has two locations in which it looks for unit files. I think this is probably the same for the target files, although there wasn’t any target files in the system administrator defined unit location but there were some target.wants files there.

systemd Monit Unit file

I found a template that Monit had already provided for a unit file in /usr/share/doc/monit/examples/monit.service. There’s also one for Upstart. Copy that to where the system administrator unit files should go and make the change so that systemd restarts Monit if it dies for what ever reason. Check the Restart= options on the systemd.service man page. The following is what my initial unit file looked like:

[Unit]
Description=Pro-active monitoring utility for unix systems
After=network.target

[Service]
Type=simple
ExecStart=/usr/bin/monit -I -c /etc/monit/monitrc
ExecStop=/usr/bin/monit -c /etc/monit/monitrc quit
ExecReload=/usr/bin/monit -c /etc/monit/monitrc reload
Restart=always

[Install]
WantedBy=multi-user.target

Now, some explanation. Most of this is pretty obvious. The After= directive just tells systemd to make sure the network.target file has been acted on first and of course network.target has After=network-pre.target which doesn’t have a lot in it. I’m not going to go into this now, as I don’t really care too much about it. It works. It means the network interfaces have to be up first. If you want to know how, why, check this documentation. Type=simple. Again check the systemd.service man page.
Now to have systemd control Monit, Monit must not run as a background process (the default). To do this, we can either add the set init statement to Monit’s control file or add the -I option when running systemd, which is exactly what we’ve done above. The WantedBy= is the target that this specific unit is part of.

Now we need to tell systemd to create the symlinks in multi-user.target.wants directory and other things. See the man page for more details about what enable actually does if you want them. You’ll also need to start the unit.

Now what I like to do here is:

systemctl status /etc/systemd/system/monit.service

Then compare this output once we enable the service:

● monit.service - Pro-active monitoring utility for unix systems
   Loaded: loaded (/etc/systemd/system/monit.service; disabled)
   Active: inactive (dead)
sudo systemctl enable /etc/systemd/system/monit.service
# systemd now knows about monit.service
systemctl status /etc/systemd/system/monit.service

Outputs:

● monit.service - Pro-active monitoring utility for unix systems
   Loaded: loaded (/etc/systemd/system/monit.service; enabled)
   Active: inactive (dead)

Now start the service:

sudo systemctl start monit.service # there's a stop and restart also.

Now you can check the status of your Monit service again. This shows terse runtime information about the units or PID you specify (monit.service in our case).

sudo systemctl status monit.service

By default this function will show you 10 lines of output. The number of lines can be controlled with the --lines= option

sudo systemctl --lines=20 status monit.service

Now try killing the Monit process. At the same time, you can watch the output of Monit in another terminal. tmux or screen is helpful for this:

sudo tail -f /var/log/monit.log
sudo kill -SIGTERM $(pidof monit)
# SIGTERM is a safe kill and is the default, so you don't actually need to specify it. Be patient, this may take a minute or two for the Monit process to terminate.

Or you can emulate a nastier termination with SIGKILL or even SEGV (which may kill monit faster).

Now when you run another status command you should see the PID has changed. This is because systemd has restarted Monit.

When you need to make modifications to the unit file, you’ll need to run the following command after save:

sudo systemctl daemon-reload

When you need to make modifications to the running services configuration file
/etc/monit/monitrc for example, you’ll need to run the following command after save:

sudo systemctl reload monit.service
# because systemd is now in control of Monit, rather than the before mentioned: sudo monit reload

 

Keep NodeJS Application Alive

Right, we know systemd is always going to be running. So lets use it to take care of the coarse grained service control. That is keeping your NodeJS application service alive.

Using systemd

systemd my-web-app.service Unit file

You’ll need to know where your NodeJS binary is. The following will provide the path:

which NodeJS

Now create a systemd unit file my-nodejs-app.service

[Unit]
Description=My amazing NodeJS application
After=network.target

[Service]
# systemctl start my-nodejs-app # to start the NodeJS script
ExecStart=[where nodejs binary lives] [where your app.js/index.js lives]
# systemctl stop my-nodejs-app # to stop the NodeJS script
# SIGTERM (15) - Termination signal. This is the default and safest way to kill process.
# SIGKILL (9) - Kill signal. Use SIGKILL as a last resort to kill process. This will not save data or cleaning kill the process.
ExecStop=/bin/kill -SIGTERM $MAINPID
# systemctl reload my-nodejs-app # to perform a zero-downtime restart.
# SIGHUP (1) - Hangup detected on controlling terminal or death of controlling process. Use SIGHUP to reload configuration files and open/close log files.
ExecReload=/bin/kill -HUP $MAINPID
Restart=always
StandardOutput=syslog
StandardError=syslog
SyslogIdentifier=my-nodejs-app
User=my-nodejs-app
Group=my-nodejs-app # Not really needed unless it's different, as the default group of the user is chosen without this option. Self documenting though, so I like to have it present.
Environment=NODE_ENV=production

[Install]
WantedBy=multi-user.target

Add the system user and group so systemd can actually run your service as the user you’ve specified.

sudo groupadd --system my-nodejs-app # this is not needed if you adduser like below...
getent group # to verify which groups exist.
sudo adduser --system --no-create-home --group my-nodejs-app # This will create a system group with the same name and ID of the user.
groups my-nodejs-app # to verify which groups the new user is in.

Now as we did above, go through the same procedure enable‘ing, start‘ing and verifying your new service.

Make sure you have your directory permissions set-up correctly and you should have a running NodeJS application that when it dies will be restarted automatically by systemd.

Don’t forget to backup all your new files and changes in case something happens to your server.

We’re done with systemd for now. Following are some useful resources I’ve used:

 

Using Monit

Now just configure your Monit control file. You can spend a lot of time here tweaking a lot more than just your NodeJS application. There are loads of examples around and the control file itself has lots of commented out examples also. You’ll find the following the most helpful:

There are a few things that had me stuck for a bit. By default Monit only sends alerts on change, not on every cycle if the condition stays the same, unless when you set-up your

set alert your-ame@your.domain

Append receive all alerts, so that it looks like this:

set alert your-ame@your.domain receive all alerts

There’s quite a few things you just work out as you go. The main part I used to health-check my NodeJS app was:

check host myhost with address 1.2.3.4
   start program = "/bin/systemctl start my-nodejs-app.service"
   stop program = "/bin/systemctl stop my-nodejs-app.service"
   if failed ping then alert
   if failed
      port 80 and
      protocol http and
      status = 200 # The default without status is failure if status code >= 400
      request /testdir with content = "some text on my web page" and
         then restart
   if 5 restarts within 5 cycles then alert

I carry on and check things like:

  1. cpu and memory usage
  2. load averages
  3. File system space on all the mount points
  4. Check SSH that it hasn’t been restarted by anything other than Monit (potentially swapping the binary or it’s config). Of course if an attacker kills Monit, systemd immediately restarts it and we get Monit alert(s). We also get real-time logging hopefully to an off-site syslog server. Ideally your off-site syslog server also has alerts set-up on particular log events. On top of that you should also have inactivity alerts set-up so that if your log files are not generating events that you expect, then you also receive alerts. Services like Dead Man’s Snitch or packages like Simple Event Correlator with Cron are good for this. On top of all that, if you have a file integrity checker that resides on another system that your host reveals no details of and you’ve got it configured to check all the right file check-sums, dates, permissions, etc, you’re removing a lot of low hanging fruit for someone wanting to compromise your system.
  5. Directory permissions, uid, gid and checksums. Of course you’re also going to have to make sure the tools that Monit uses to do these checks haven’t been modified.

 

If you find anything I haven’t explained clearly, or you need a hand with any of this just leave a comment. Cheers.

Advertisements

Evaluation of Host Intrusion Detection Systems (HIDS)

May 30, 2015

Followed up with a test deployment and drive.

The best time to install a HIDS is on a fresh install before you open the host up to the internet or even your LAN if it’s corporate. Of course if you don’t have that luxury, there are a bunch of tools that can help you determine if you’re already owned. Be sure to run one or more over your target system before your HIDS bench-marks it.

The reason I chose Stealth and OSSEC to take further into an evaluation was because they rose to the top of a preliminary evaluation I performed during a recent Debian Web server hardening process where I looked at several other HIDS as well.

OSSEC

ossec-hids

Source

ossec-hids on github

Who’s Behind Ossec

  • Many developers, contributors, managers, reviewers, translators (infact the OSSEC team looks almost as large as the Stealth user base)

Documentation

Lots of documentation. Not always the easiest to navigate. Lots of buzz on the inter-webs.
Several books.

Community / Communication

IRC channel #ossec at irc.freenode.org Although it’s not very active.

Components

  • Manager (sometimes called server): does most of the work monitoring the Agents. It stores the file integrity checking databases, the logs, events and system auditing entries, rules, decoders, major configuration options.
  • Agents: small collections of programs installed on the machines we are interested in monitoring. Agents collect information and forward it to the manager for analysis and correlation.

There are quite a few other ancillary components also.

Architecture

You can also go the agent-less route which may allow the Manager to perform file integrity checks using agent-less scripts. As with Stealth, you’ve still got the issue of needing to be root in order to read some of the files.

Agents can be installed on VMware ESX but from what I’ve read it’s quite a bit of work.

Features (What does it do)

  • File integrity checking
  • Rootkit detection
  • Real-time log file monitoring and analysis (you may already have something else doing this)
  • Intrusion Prevention System (IPS) features as well: blocking attacks in real-time
  • Alerts can go to a databases MySQL or PostgreSQL or other types of outputs
  • There is a PHP web UI that runs on Apache if you would rather look at pretty outputs vs log files.

What I like

To me, the ability to scan in real-time off-sets the fact that the agents need binaries installed. This hinders the attacker from covering their tracks.

Can be configured to scan systems in realtime based on inotify events.

Backed by a large company Trend Micro.

Options. Install options for starters. You’ve got the options of:

  • Agent-less installation as described above
  • Local installation: Used to secure and protect a single host
  • Agent installation: Used to secure and protect hosts while reporting back to a
    central OSSEC server
  • Server installation: Used to aggregate information

Can install a web UI on the manager, so you need Apache, PHP, MySQL.

What I don’t like

  • Unlike Stealth, The fact that something has to be installed on the agents
  • The packages are not in the standard repositories. The downloads, PGP keys and directions are here.
  • I think Ossec may be doing to much and if you don’t like the way it does one thing, you may be stuck with it. Personally I really like the idea of a tool doing one thing, doing it well and providing plenty of configuration options to change the way it does it’s one thing. This provides huge flexibility and minimises your dependency on a suite of tools and/or libraries
  • Information overload. There seems to be a lot to get your head around to get it set-up. There are a lot of install options documented (books, interwebs, official docs). It takes a bit to workout exactly the best procedure for your environment. Following are some of the resources I used:
    1. Some official documentation
    2. Perving in the repository
    3. User take one install
    4. User take two install
    5. User take three install

Stealth

Stealth file integrity checker

Source

sourceforge

Who’s Behind Stealth

Author: Frank B. Brokken. An admirable job for one person. Frank is not a fly-by-nighter though. Stealth was first presented to Congress in 2003. It’s still actively maintained and used by a few. It’s one of GNU/Linux’s dirty little secrets I think. It’s a great idea, makes a tricky job simple and does it in an elegant way.

Documentation

  • 3.00.00 (2014-08-29)
  • 2.11.03 (2013-06-18)
    • Once you install Stealth, all the documentation can be found by sudo updatedb && locate stealth. I most commonly used: HTML docs (/usr/share/doc/stealth-doc/manual/html/) and (/usr/share/doc/stealth-doc/manual/pdf/stealth.pdf) for easy searching across the HTML docs
    • man page (/usr/share/doc/stealth/stealthman.html)
    • Examples: (/usr/share/doc/stealth/examples/)
  • More covered in my demo set-up below

Binaries

  • 3.00.00 (2014-08-29) is in the repository for Debain Jessie
  • 2.11.03-2 (2013-06-18) This is as recent as you can go for Linux Mint Qiana (17) and Rebecca (17.1) within the repositories, unless you want to go out of band. There are quite a few dependencies ldd will show you:
    libbobcat.so.3 => /usr/lib/libbobcat.so.3 (0xb765d000)
    libpthread.so.0 => /lib/i386-linux-gnu/i686/cmov/libpthread.so.0 (0xb7641000)
    libstdc++.so.6 => /usr/lib/i386-linux-gnu/libstdc++.so.6 (0xb754e000)
    libm.so.6 => /lib/i386-linux-gnu/i686/cmov/libm.so.6 (0xb7508000)
    libgcc_s.so.1 => /lib/i386-linux-gnu/libgcc_s.so.1 (0xb74eb000)
    libc.so.6 => /lib/i386-linux-gnu/i686/cmov/libc.so.6 (0xb7340000)
    libX11.so.6 => /usr/lib/i386-linux-gnu/libX11.so.6 (0xb71ee000)
    libcrypto.so.1.0.0 => /usr/lib/i386-linux-gnu/i686/cmov/libcrypto.so.1.0.0 (0xb7022000)
    libreadline.so.6 => /lib/i386-linux-gnu/libreadline.so.6 (0xb6fdc000)
    libmilter.so.1.0.1 => /usr/lib/i386-linux-gnu/libmilter.so.1.0.1 (0xb6fcb000)
    /lib/ld-linux.so.2 (0xb7748000)
    libxcb.so.1 => /usr/lib/i386-linux-gnu/libxcb.so.1 (0xb6fa5000)
    libdl.so.2 => /lib/i386-linux-gnu/i686/cmov/libdl.so.2 (0xb6fa0000)
    libtinfo.so.5 => /lib/i386-linux-gnu/libtinfo.so.5 (0xb6f7c000)
    libXau.so.6 => /usr/lib/i386-linux-gnu/libXau.so.6 (0xb6f78000)
    libXdmcp.so.6 => /usr/lib/i386-linux-gnu/libXdmcp.so.6 (0xb6f72000)
    
  • 2.10.01 (2012-10-04) is in the repository for Debian Wheezy

Community / Communication

No community really. I see it as one of the dirty little secretes that I’m surprised many diligent sysadmins haven’t jumped on. The author is happy to answer emails. The author doesn’t market.

Components

Controller

The computer initiating the scan.

Needs two kinds of outgoing services:

  1. ssh to reach the clients
  2. mail transport agent (MTA)(sendmail, postfix)

Considerations for the controller:

  1. No public access.
  2. All inbound services should be denied.
  3. Access only via its console.
  4. Physically secure location (one would think goes without saying, but you may be surprised).
  5. Sensitive information of the clients are stored on the controller.
  6. Password-less access to the clients for anyone who gains controller root access, unless ssh-cron is used, which appears to have been around since 2014-05.

Client

The computer/s being scanned. I don’t see any reason why a Stealth solution couldn’t be set-up to look after many clients.

Architecture

The controller stores one to many policy files. Each of which is specific to a single client and contains use directives and commands. It’s recommended policy to take copies of the client programmes such as the hashing programme sha1sum, find and others that are used extensively during the integrity scans and copy them to the controller to take bench-mark hashes. Subsequent runs will do the same to compare with the initial hashes stored.

Features (What does it do)

File integrity tests leaving virtually no sediments on the tested client.

Stealth subscribes to the “dark cockpit” approach. I.E. no mail is sent when no changes were detected. If you have a MTA, Stealth can be configured to send emails on changes it finds.

What I like

  • It’s simplicity. There is one package to install on the controller. Nothing to install on the client machines. Client just needs to have the controllers SSH public key. You will need a Mail Transfer Agent on your controller if you don’t already have one. My test machine (Linux Mint) didn’t have one.
  • Rather than just modifying the likes of sha1sum on the clients that Stealth uses to perform it’s integrity checks, Stealth would somehow have to be fooled into thinking that the changed hash of the sha1sum it’s just copied to the controller is the same as the previously recorded hash that it did the same with. If the previously recorded hash is removed or does not match the current hash, then Stealth will fire an alert off.
  • It’s in the Debian repositories. Which is a little surprising considering I couldn’t find any test suite results.
  • The whole idea behind it. Systems being monitored give little appearance that they’re being monitored, other than I think the presence of a single SSH login when Stealth first starts in the auth.log. This could actually be months ago, as the connection remains active for the life of Stealth. The login could be from a user doing anything on the client. It’s very discrete.
  • unpredictability of Stealth runs is offered through Stealth’s --random-interval and --repeat options. E.g., --repeat 60 --random-interval 30 results in new Stealth-runs on average every 75 seconds.
  • Subscribes to the Unix philosophy: “do one thing and do it well”
  • Stealth’s author is very approachable and open. After talking with Frank and suggesting some ideas to promote Stealth and it’s community, Frank started a discussion list.

What I don’t like

  • Lack of visible code reviews and testing.Yes it’s in Debian, but so was OpenSSL and Bash
  • One man band. Support provided via one person alone via email. Comparing with the likes of Ossec which has …
  • Lack of use cases. I don’t see anyone using/abusing it. Although Frank did send me some contacts of other people that are using it, so again, a very helpful author. Can’t find much on the interwebs. The documentation has clear signs that it’s been written and is targeting people already familiar with the tool. This is understandable as the author has been working on this project for nine years and could possibly be disconnected with what’s involved for someone completely new to the project to dive in and start using it. In saying that, that’s what I did and so far it worked out well.
  • This tells me that either very few are using it or it has no bugs and the install and configuration is stupidly straight forward or both.
  • Small user base. This is how many people are happy to reveal they are using Stealth.
  • Reading through the userguide, the following put me off a little: “preferably the public ssh-identity key of the controller should be placed in the client’s root .ssh/authorized_keys file, granting the controller root access to the client. Root access is normally needed to gain access to all directories and files of the client’s file system.” I never allow SSH root access to servers. So I’m not about to start. What’s worse, is that Stealth SSHing from server to client with key-pair can only do so automatically if the pass-phrase is blank. If it’s not blank then someone has to drive Stealth and the whole idea of Stealth (as far as I can tell) is to be an automatic file integrity checker.
    There are however some work-arounds to this. ssh-cron can run scheduled Stealth jobs, but needs to aquire the pass-phrase from the user once. It does this via ssh-askpass which is a X11 based pass-phrase input dialog. If your running ssh-cron and Stealth from a server (which would be a large number of potential users I would think) you won’t have X11. So if that’s the case ssh-cron is out of the question. At least that’s how I understand it from the man page and emails from Frank. Frank mentioned in an email: “Stealth itself could perfectly well be started `by hand’ setting up the connection using, e.g., an existing ssh private-key, which could thereafter completely be removed from the system running Stealth. The running Stealth process would then continue to use the established connection for as long as it’s running. This may be a very long time if the –daemon option is used.” I’ve used Monit to do any checks on the client that need root access. This way Stealth can run as a low privileged user.
  • In order to use an SSH key-pair with passphrase and have your controller resume scans on reboot, you’re going to need ssh-cron. Some distros like Mint and Ubuntu only have very old versions of libbobcat (3.19.01 (December 2013)) in their repositories. You could re-build if you fancy the dependency issues it may bring. Failing that, use Debian which is way more up to date, or just fire stealth off each time you reboot your controller machine manually and run it as a daemon with arguments such as --keep-alive (or --daemon if running stealth >= 4.00.00), --repeat. This will cause Stealth to keep running (sleeping) most of the time, then doing it’s checks at the intervals you define.

Outcomes

In making all of my considerations, I changed my mind quite a few times on which offerings were most suited to which environments. I think this is actually a good thing, as I think it means my evaluations were based on the real merits of each offering rather than any biases.

If you already have real-time logging to an off-site syslog server set-up and alerting. OSSEC would provide redundant features.

If you don’t already have real-time logging to an off-site syslog server, then OSSEC can help here.

The simplicity of Stealth, flatter learning curve and it’s over-all philosophy is what won me over.

Stealth Up and Running

I tested this out on a Mint installation.

Installed stealth and stealth-doc (2.11.03-2) via synaptic package manager. Then just did a locate for stealth to find the docs and other example files. The following are the files I used for documentation, how I used them and the tab order that made sense to me:

  1. The main documentation index: file:///usr/share/doc/stealth-doc/manual/html/stealth.html
  2. Chapter one introduction: file:///usr/share/doc/stealth-doc/manual/html/stealth01.html
  3. Chapter three to help build up a policy file: file:///usr/share/doc/stealth-doc/manual/html/stealth03.html
  4. Chapter five for running Stealth and building up policy file: file:///usr/share/doc/stealth-doc/manual/html/stealth05.html
  5. Chapter six for running Stealth: file:///usr/share/doc/stealth-doc/manual/html/stealth06.html
  6. Chapter seven for arguements to pass to Stealth: file:///usr/share/doc/stealth-doc/manual/html/stealth07.html
  7. Chapter eight for error messages: file:///usr/share/doc/stealth-doc/manual/html/stealth08.html
  8. The Man page: file:///usr/share/doc/stealth/stealthman.html
  9. Policy file examples: file:///usr/share/doc/stealth/examples/
  10. Useful scripts to use with Stealth: file:///usr/share/doc/stealth/scripts/usr/bin/
  11. All of the documentation in simple text format (good for searching across chapters for strings): file:///usr/share/doc/stealth-doc/manual/text/stealth.txt

Files I would need to copy and modify were:

  • /usr/share/doc/stealth/scripts/usr/bin/stealthcleanup.gz
  • /usr/share/doc/stealth/scripts/usr/bin/stealthcron.gz
  • /usr/share/doc/stealth/scripts/usr/bin/stealthmail.gz

Files I used for reference to build up a policy file:

  • /usr/share/doc/stealth/examples/demo.pol.gz
  • /usr/share/doc/stealth/examples/localhost.pol.gz
  • /usr/share/doc/stealth/examples/simple.pol.gz

As mentioned above, providing you have a working MTA, then Stealth will just do it’s thing when you run it. The next step is to schedule it’s runs. This can be also (as mentioned above) with a pseudo random interval.

Feel free to leave a comment if you need help setting Stealth up, as it did take a bit of fiddling, but does what it says it does on the box very well.

Web Server Log Management

April 25, 2015

As part of the ongoing work around preparing a Debian web server to host applications accessible from the WWW I performed some research, analysis, made decisions along the way and implemented a first stage logging strategy. I’ve done similar set-ups many times before, but thought it worth sharing my experience for all to learn something from it and/or provide input, recommendations, corrections to the process so we all get to improve.

The main system loggers I looked into

  • GNU syslogd which I don’t think is being developed anymore? Correct me if I’m wrong. Most Linux distributions no longer ship with this. Only supports UDP. It’s also a bit lacking in features. From what I gather is single-threaded. I didn’t spend long looking at this as there wasn’t much point. The following two offerings are the main players.
  • rsyslog: which ships with Debian and most other Linux distros now I believe. I like to do as little as possible and rsyslog fits this description for me. The documentation seems pretty good. Rainer Gerhards wrote rsyslog and his blog provides some good insights. Supports UDP, TCP. Can send over TLS. There is also the Reliable Event Logging Protocol (RELP) which Rainer created.
    rsyslog is great at gathering, transporting, storing log messages and includes some really neat functionality for dividing the logs. It’s not designed to alert on logs. That’s where the likes of Simple Event Correlator (SEC) comes in. Rainer discusses why TCP isn’t as reliable as many think here.
  • syslog-ng: I didn’t spend to long here, as I didn’t see any features that I needed that were better than the default of rsyslog. Can correlate log messages, both real-time and off-line. Supports reliable and encrypted transport using TCP and TLS. message filtering, sorting, pre-processing, log normalisation.

There are are few comparisons around. Most of the ones I’ve seen are a bit biased and often out of date.

Aims

  • Record events and have them securely transferred to another syslog server in real-time, or as close to it as possible, so that potential attackers don’t have time to modify them on the local system before they’re replicated to another location
  • Reliability (resilience / ability to recover connectivity)
  • Extensibility: ability to add more machines and be able to aggregate events from many sources on many machines
  • Receive notifications from the upstream syslog server of specific events. No HIDS is going to remove the need to reinstall your system if you are not notified in time and an attacker plants and activates their root-kit.
  • Receive notifications from the upstream syslog server of lack of events. The network is down for example.

Environmental Considerations

A couple of servers in the mix:

FreeNAS File Server

Recent versions can send their syslog events to a syslog server. With some work, it looks like FreeNAS can be setup to act as a syslog server.

pfSense Router

Can send log events, but only by UDP by the look of it.

Following are the two strategies that emerged. You can see by the detail that I went down the path of the first one initially. It was the path of least resistance / quickest to setup. I’m going to be moving away from papertrail toward strategy two. Mainly because I’ve had a few issues where messages have been getting lost that have been very hard to track down (I’ve spent over a week on it). As the sender, you have no insight into what papertrail is doing. The support team don’t provide a lot of insight into their service when you have to trouble-shoot things. They have been as helpful as they can be, but I’ve expressed concern around them being unable to trouble-shoot their own services.

Outcomes

Strategy One

Rsyslog, TCP, local queuing, TLS, papertrail for your syslog server (PT doesn’t support RELP, but say that’s because their clients haven’t seen any issues with reliability in using plain TCP over TLS with local queuing). My guess is they haven’t looked hard enough. I must be the first then. Beware!

As I was setting this up and watching both ends. We had an internet outage of just over an hour. At that stage we had very few events being generated, so it was trivial to verify both ends. I noticed that once the ISP’s router was back on-line and the events from the queue moved to papertrail, that there was in fact one missing.

Why did Rainer Gerhards create RELP if TCP with queues was good enough? That was a question that was playing on me for a while. In the end, it was obvious that TCP without RELP isn’t good enough.
At this stage it looks like the queues may loose messages. Rainer says things like “In rsyslog, every action runs on its own queue and each queue can be set to buffer data if the action is not ready. Of course, you must be able to detect that the action is not ready, which means the remote server is off-line. This can be detected with plain TCP syslog and RELP“, but it can be detected without RELP.

You can aggregate log files with rsyslog or by using papertrails remote_syslog daemon.

Alerting is available, including for inactivity of events.

Papertrails documentation is good and support is reasonable. Due to the huge amounts of traffic they have to deal with, they are unable to trouble-shoot any issues you may have. If you still want to go down the papertrail path, to get started, work through this which sets up your rsyslog to use UDP (specified in the /etc/rsyslog.conf by a single ampersand in front of the target syslog server). I want something more reliable than that, so I use two ampersands, which specifies TCP.

As we’re going to be sending our logs over the internet for now, we need TLS. Check papertrails CA server bundle for integrity:

curl https://papertrailapp.com/tools/papertrail-bundle.pem | md5sum

Should be: c75ce425e553e416bde4e412439e3d09

If all good throw the contents of that URL into a file called papertrail-bundle.pem.
Then scp the papertrail-bundle.pem into the web servers /etc dir. The command for that will depend on whether you’re already on the web server and you want to pull, or whether you’re somewhere else and want to push. Then make sure the ownership is correct on the pem file.

chown root:root papertrail-bundle.pem

install rsyslog-gnutls

apt-get install rsyslog-gnutls

Add the TLS config

$DefaultNetstreamDriverCAFile /etc/papertrail-bundle.pem # trust these CAs
$ActionSendStreamDriver gtls # use gtls netstream driver
$ActionSendStreamDriverMode 1 # require TLS
$ActionSendStreamDriverAuthMode x509/name # authenticate by host-name
$ActionSendStreamDriverPermittedPeer *.papertrailapp.com

to your /etc/rsyslog.conf. Create egress rule for your router to let traffic out to dest port 39871.

sudo service rsyslog restart

To generate a log message that uses your system syslogd config /etc/rsyslog.conf, run:

logger "hi"

should log “hi” to /var/log/messages and also to papertrail, but it wasn’t.

# Show a live update of the last 10 lines (by default) of /var/log/messages
sudo tail -f [-n <number of lines to tail>] /var/log/messages

OK, so lets run rsyslog in config checking mode:

/usr/sbin/rsyslogd -f /etc/rsyslog.conf -N1

Output all good looks like:

rsyslogd: version <the version number>, config validation run (level 1), master config /etc/rsyslog.conf
rsyslogd: End of config validation run. Bye.

Trouble-shooting

  1. https://www.loggly.com/docs/troubleshooting-rsyslog/
  2. http://help.papertrailapp.com/
  3. http://help.papertrailapp.com/kb/configuration/troubleshooting-remote-syslog-reachability/
  4. /usr/sbin/rsyslogd -version will provide the installed version and supported features.

Which didn’t help a lot, as I don’t have telnet installed. I can’t ping from the DMZ as ICMP is not allowed out and I’m not going to install tcpdump or strace on a production server. The more you have running, the more surface area you have, the greater the opportunities to exploit.

So how do we tell if rsyslogd is actually running if it doesn’t appear to be doing anything useful?

pidof rsyslogd

or

/etc/init.d/rsyslog status

Showing which files rsyslogd has open can be useful:

lsof -p <rsyslogd pid>

or just combine the results of pidof rsyslogd

sudo lsof -p $(pidof rsyslogd)

To start with I had a line like:

rsyslogd 3426 root 8u IPv4 9636 0t0 TCP <web server IP>:<sending port>->logs2.papertrailapp.com:39871 (SYN_SENT)

Which obviously showed rsyslogd‘s SYN packets were not getting through. I’ve had some discussion with Troy from PT support around the reliability of plain TCP over TLS without RELP. I think if the server is business critical, then strategy two “maybe” the better option. Troy has assured me that they’ve never had any issues with logs being lost due to lack of reliability with out RELP. Troy also pointed me to their recommended local queue options. After adding the queue tweaks and a rsyslogd restart, it resulted in:

rsyslogd 3615 root 8u IPv4 9766 0t0 TCP <web server IP>:<sending port>->logs2.papertrailapp.com:39871 (ESTABLISHED)

I could now see events in the papertrail web UI in real-time.

Socket Statistics (ss)(the better netstat) should also show the established connection.

By default papertrail accepts TCP over TLS (TLS encryption check-box on, Plain text check-box off) and UDP. So if your TLS isn’t setup properly, your events won’t be accepted by papertrail. I later confirmed this to be true.

Check that our Logs are Commuting over TLS

Now without installing anything on the web server or router, or physically touching the server sending packets to papertrail or the router. Using a switch (ubiquitous) rather than a hub. No wire tap or multi-network interfaced computer. No switch monitoring port available on expensive enterprise grade switches (along with the much needed access). We’re basically down to two approaches I can think of and I really couldn’t be bothered getting up out of my chair.

  1. MAC flooding with the help of macof which is a utility from the dsniff suite. This essentially causes your switch to go into a “failopen mode” where it acts like a hub and broadcasts it’s packets to every port.

    MAC Flooding

    Or…
  2. Man in the Middle (MiTM) with some help from ARP spoofing or poisoning. I decided to choose the second option, as it’s a little more elegant.

    ARP Spoofing

On our MitM box, I set a static IP: address, netmask, gateway in /etc/network/interfaces and add domain, search and nameservers to the /etc/resolv.conf.

Follow that up with a service network-manager restart

On the web server run:

ifconfig -a

to get MAC: <MitM box MAC> On MitM box run the same command to get MAC: <web server MAC>
On web server run:

ip neighbour

to find MACs associated with IP’s (the local ARP table). Router was: <router MAC>.

myuser@webserver:~$ ip neighbour
<MitM box IP> dev eth0 lladdr <MitM box MAC> REACHABLE
<router IP> dev eth0 lladdr <router MAC> REACHABLE

Now you need to turn your MitM box into a router temporarily. On the MitM box run

cat /proc/sys/net/ipv4/ip_forward

You’ll see a ‘1’ if forwarding is on. If it’s not, throw a ‘1’ into the file:

echo 1 > /proc/sys/net/ipv4/ip_forward

and check again to make sure. Now on the MitM box run

arpspoof -t <web server IP> <router IP>

This will continue to notify <web server IP> that our (MitM box) MAC address belongs to <router IP>. Essentially… we (MitM box) are <router IP> to the <web server IP> box, but our IP address doesn’t change. Now on the web server you can see that it’s ARP table has been updated and because arpspoof keeps running, it keeps telling <web server IP> that our MitM box is the router.

myuser@webserver:~$ ip neighbour
<MitM box IP> dev eth0 lladdr <MitM box MAC> STALE
<router IP> dev eth0 lladdr <MitM box MAC> REACHABLE

Now on our MitM box, while our arpspoof continues to run, we start Wireshark listening on our eth0 interface or what ever interface your using, and you can see that all packets that the web server is sending, we are intercepting and forwarding (routing) on to the gateway.

Now Wireshark clearly showed that the data was encrypted. I commented out the five TLS config lines in the /etc/rsyslog.conf file -> saved -> restarted rsyslog -> turned on “Plain text” in papertrail and could now see the messages in clear text. Now when I turned off “Plain text” papertrail would no longer accept syslog events. Excellent!

One of the nice things about arpspoof is that it re-applies the original ARP’s once it’s done.

You can also tell arpspoof to poison the routers ARP table. This way any traffic going to the web server via the router, not originating from the web server will be routed through our MitM box also.

Don’t forget to revert the change to /proc/sys/net/ipv4/ip_forward.

Exporting Wireshark Capture

You can use the File->Save As… option here for a collection of output types, or the way I usually do it is:

  1. First completely expand all the frames you want visible in your capture file
  2. File->Export Packet Dissections->as “Plain Text” file…
  3. Check the “All packets” check-box
  4. Check the “Packet summary line” check-box
  5. Check the “Packet details:” check-box and the “As displayed”
  6. OK

Trouble-shooting messages that papertrail never shows

To run rsyslogd in debug

Check to see which arguments get passed into rsyslogd to run as a daemon in /etc/init.d/rsyslog and /etc/default/rsyslog. You’ll probably see a RSYSLOGD_OPTIONS="". There may be some arguments between the quotes.

sudo service rsyslog stop
sudo /usr/sbin/rsyslogd [your options here] -dn >> ~/rsyslog-debug.log

The debug log can be quite useful for trouble-shooting. Also keep your eye on the stderr as you can see if it’s writing anything out (most system start-up scripts throw this away).
Once you’ve finished collecting log:
ctrl+C

sudo service rsyslog start

To see if rsyslog is running

pidof rsyslogd
# or
/etc/init.d/rsyslog status
Turn on the impstats module

The stats it produces show when you run into errors with an output, and also the state of the queues.
You can also run impstats on the receiving machine if it’s in your control. Papertrail obviously is not.
Put the following into your rsyslog.conf file at the top and restart rsyslog:

# Turn on some internal counters to trouble-shoot missing messages
module(load="impstats"
interval="600"
severity="7"
log.syslog="off"

# need to turn log stream logging off
log.file="/var/log/rsyslog-stats.log")
# End turn on some internal counters to trouble-shoot missing messages

Now if you get an error like:

rsyslogd-2039: Could not open output pipe '/dev/xconsole': No such file or directory [try http://www.rsyslog.com/e/2039 ]

You can just change the /dev/xconsole to /dev/console
xconsole is still in the config file for legacy reasons, it should have been cleaned up by the package maintainers.

GnuTLS error in rsyslog-debug.log

By running rsyslogd manually in debug mode, I found an error when the message failed to send:

unexpected GnuTLS error -53 in nsd_gtls.c:1571

Standard Error when running rsyslogd manually produces:

GnuTLS error: Error in the push function

With some help from the GnuTLS mailing list:

That means that send() returned -1 for some reason.” You can enable more output by adding an environment variable GNUTLS_DEBUG_LEVEL=9 prior to running the application, and that should at least provide you with the errno. This didn’t actually provide any more detail to stderr. However, thanks to Rainer we do now have debug.gnutls parameter in the rsyslog code that if you specify this global variable in the rsyslog.conf and assign it a value between 0-10 you’ll have gnutls debug output going to rsyslog’s debug log.

Strategy Two

Rsyslog, TCP, local queuing, TLS, RELP, SEC, syslog server on local network. Notification for inactivity of events could be performed by cron and SEC?
LogAnalyzer also created by Rainer Gerhards (rsyslog author), but more work to setup than an on-line service you don’t have to setup. In saying that. You would have greater control and security which for me is the big win here.
Normalisation also looks like Rainer has his finger in this pie.

In theory Adding RELP to TCP with local queues is a step-up in terms of reliability. Others have said, the reliability of TCP over TLS with local queues is excellent anyway. I’ve yet to confirm it’s excellence. At the time of writing this post,I’m seriously considering moving toward RELP to help solve my reliability issues.

Additional Resource

gentoo rsyslog wiki

Keeping Your Linux Server/s In Time With Your Router

March 28, 2015

Your NTP Server

With this set-up, we’ve got one-to-many Linux servers in a network that all want to be synced with the same up-stream Network Time Protocol (NTP) server/s that your router (or what ever server you choose to be your NTP authority) uses.

On your router or what ever your NTP server host is, add the NTP server pools. Now how you do this really depends on what your using for your NTP server, so I’ll leave this part out of scope. There are many NTP pools you can choose from. Pick one or a collection that’s as close to you’re NTP server as possible.

If your NTP daemon is running on your router, you’ll need to decide and select which router interfaces you want the NTP daemon supplying time to. You almost certainly won’t want it on the WAN interface (unless you’re a pool member) if you have one on your router.

Make sure you restart your NTP daemon.

Your Client Machines

If you have ntpdate installed, /etc/default/ntpdate says to look at /etc/ntp.conf which doesn’t exist without ntp being installed. It looks like this:

# Set to "yes" to take the server list from /etc/ntp.conf, from package ntp,
# so you only have to keep it in one place.
NTPDATE_USE_NTP_CONF=yes

but you’ll see that it also has a default NTPSERVERS variable set which is overridden if you add your time server to /etc/ntp.conf. If you enter the following and ntpdate is installed:

dpkg-query -W -f='${Status} ${Version}\n' ntpdate

You’ll get output like:

install ok installed 1:4.2.6.p5+dfsg-3ubuntu2

Otherwise install it:

apt-get install ntp

The public NTP server/s can be added straight to the bottom of the /etc/ntp.conf file, but because we want to use our own NTP server, we add the IP address of our server that’s configured with our NTP pools to the bottom of the file.

server <IP address of your local NTP server here>

Now if your NTP daemon is running on your router, hopefully you have everything blocked on its interface/s by default and are using a white-list for egress filtering.

In which case you’ll need to add a firewall rule to each interface of the router that you want NTP served up on.

NTP talks over UDP and listens on port 123 by default.

After any configuration changes to your ntpd make sure you restart it. On most routers this is done via the web UI.

On the client (Linux) machines:

sudo service ntp restart

Now issuing the date command on your Linux machine will provide the current time, yes with seconds.

Trouble-shooting

The main two commands I use are:

sudo ntpq -c lpeer

Which should produce output like:

            remote                       refid         st t when poll reach delay offset jitter
===============================================================================================
*<server name>.<domain name> <upstream ntp ip address> 2  u  54   64   77   0.189 16.714 11.589

and the standard NTP query program followed by the as argument:

ntpq

Which will drop you at ntpq’s prompt:

ntpq> as

Which should produce output like:

ind assid status  conf reach auth condition  last_event cnt
===========================================================
  1 15720  963a   yes   yes  none  sys.peer    sys_peer  3

Now in the first output, the * in front of the remote means the server is getting it’s time successfully from the upstream NTP server/s which needs to be the case in our scenario. Often you may also get a refid of .INIT. which is one of the “Kiss-o’-Death Codes” which means “The association has not yet synchronized for the first time”. See the NTP parameters. I’ve found that sometimes you just need to be patient here.

In the second output, if you get a condition of reject, it’s usually because your local ntp can’t access the NTP server you set-up. Check your firewall rules etc.

Now check all the times are in sync with the date command.