Tips for monitoring Rancher Server

Last week I encountered an interesting bug in Rancher that managed to cause some major problems across my Rancher infrastructure.  Basically, the bug was causing of the Rancher agent clients to continuously bounce between disconnected/reconnected/finished and reconnecting states, which only manifested itself either after a 12 hour period or by deactivating/activating agents (for example adding a new host to an environment).  The only way to temporarily fix the issue was to restart the rancher-server container.

With some help, we were eventually able to resolve the issue.  I picked up a few nice lessons along the way and also became intimately acquainted with some of the inner workings of Rancher.  Through this experience I learned some tips on how to effectively monitor the Rancher server environment that I would otherwise not have been exposed to, which I would like to share with others today.

All said and done, I view this experience as a positive one.  Hitting the bug has not only helped mitigate this specific issue for other users in the future but also taught me a lot about the inner workings of Rancher.  If you’re interested in the full story you can read about all of the details about the incident, including steps to reliably reproduce and how the issue was ultimately resolved here.  It was a bug specific to Rancher v1.5.1-3, so upgrading to 1.5.4 should fix this issue if you come across it.

Before diving into the specifics for this post, I just want to give a shout out to the Rancher community, including @cjellik, @ibuildthecloud, @ecliptok and @shakefu.  The Rancher developers, team and community members were extremely friendly and helpful in addressing and fixing the issue.  Between all the late night messages in the Rancher slack, many many logs, countless hours debugging and troubleshooting I just wanted to say thank you to everyone for the help.  The small things go a long way, and it just shows how great the growing Rancher community is.

Effective monitoring

I use Sysdig as the main source of container and infrastructure monitoring.  To accomplish the metric collection, I run the Sysdig agent as a systemd service when a server starts up so when a server dies and goes away or a new one is added, Sysdig is automatically started up and begins dumping that metric data into the Sysdig Cloud for consumption through the web interface.

I have used this data to create custom dashboards which gives me a good overview about what is happening in the Rancher server environment (and others) at any given time.

sysdig dashboard

The other important thing I discovered through this process, was the role that the Rancher database plays.  For the Rancher HA setup, I am using an externally hosted RDS instance for the Rancher database and was able to fine found some interesting correlations as part of troubleshooting thanks to the metrics in Sysdig.  For example, if the database gets stressed it can cause other unintended side effects, so I set up some additional monitors and alerts for the database.

Luckily Sysdig makes the collection of these additional AWS metrics seamless.  Basically, Sysdig offers an AWS integration which pull in CloudWatch metrics and allows you to add them to dashboards and alert on them from Sysdig, which has been very nice so far.

Below are some useful metrics in helping diagnose and troubleshoot various Rancher server issues.

  • Memory usage % (server)
  • CPU % (server)
  • Heap used over time (server)
  • Number of network connections (server)
  • Network bytes by application (server)
  • Freeable memory over time (RDS)
  • Network traffic over time (RDS)

As you can see, there are quite a few things you can measure with metrics alone.  Often though, this isn’t enough to get the entire picture of what is happening in an environment.

Logs

It is also important to have access to (useful) logs in the infrastructure in order to gain insight into WHY metrics are showing up the way they do and also to help correlate log messages and errors to what exactly is going on in an environment when problems occur.  Docker has had the ability for a while now to use log drivers to customize logging, which has been helpful to us.  In the beginning, I would just SSH into the server and tail the logs with the “docker logs” command but we quickly found that to be cumbersome to do manually.

One alternative to tailing the logs manually is to configure the Docker daemon to automatically send logs to a centralized log collection system.  I use Logstash in my infrastructure with the “gelf” log driver as part of the bootstrap command that runs to start the Rancher server container, but there are other logging systems if Logstash isn’t the right fit.  Here is what the relevant configuration looks like.

...
--log-driver=gelf \
--log-opt gelf-address=udp://<logstash-server>:12201 \
--log-opt tag=rancher-server \
...

Just specify the public address of the Logstash log collector and optionally add tags.  The extra tags make filtering the logs much easier, so I definitely recommend adding at least one.

Here are a few of the Logstash filters for parsing the Rancher logs.  Be aware though, it is currently not possible to log full Java stack traces in Logstash using the gelf input.

if [tag] == "rancher-server" {
    mutate { remove_field => "command" }
    grok {
      match => [ "host", "ip-(?<ipaddr>\d{1,3}-\d{1,3}-\d{1,3}-\d{1,3})" ]
    }

    # Various filters for Rancher server log messages
    grok {
     match => [ "message", "time=\"%{TIMESTAMP_ISO8601}\" level=%{LOGLEVEL:debug_level} msg=\"%{GREEDYDATA:message_body}\"" ]
     match => [ "message", "%{TIMESTAMP_ISO8601} %{WORD:debug_level} (?<context>\[.*\]) %{GREEDYDATA:message_body}" ]
     match => [ "message", "%{DATESTAMP} http: %{WORD:http_type} %{WORD:debug_level}: %{GREEDYDATA:message_body}" ]
   }
 }

There are some issues open for addressing this, but it doesn’t seem like there is much movement on the topic, so if you see a lot of individual messages from stack traces that is the reason.

One option to mitigate the problem of stack traces would be to run a local log collection agent (in a container of course) on the rancher server host, like Filebeat or Fluentd that has the ability to clean up the logs before sending it to something like Logstash, ElasticSearch or some other centralized logging.  This approach has the added benefit of adding encryption to the logs, which GELF does not have (currently).

If you don’t have a centralized logging solution or just don’t care about rancher-server logs shipping to it – the easiest option is to tail the logs locally as I mentioned previously, using the json-file log format.  The only additional configuration I would recommend to the json-file format is to turn on log rotation which can be accomplished with the following configuration.

...
 --log-driver=json-file \
 --log-opt max-size=100mb \
 --log-opt max-file=2 \
...

Adding these logging options will ensure that the container logs for rancher-server will never full up the disk on the server.

Bonus: Debug logs

Additional debug logs can be found inside of each rancher-server container.  Since these debug logs are typically not needed in day to day operations, they are sort of an easter egg, tucked away.  To access these debug logs, they are located in /var/lib/cattle/logs/ inside of the rancher-server container.  The easiest way to analyze the logs is to get them off the server and onto a local machine.

Below is a sample of how to do this.

docker exec -it <rancher-server> bash
cd /var/lib/cattle/logs
cp cattle-debug.log /tmp

Then from the host that the container is sitting on you can docker cp the logs out of the container and onto the working directory of the host.

docker cp <rancher-server>:/tmp/cattle-debug.log .

From here you can either analyze the logs in a text editor available on the server, or you can copy the logs over to a local machine.  In the example below, the server uses ssh keys for authentication and I chose to copy the logs from the server into my local /tmp directory.

 scp -i ~/.ssh/<rancher-server-pem> user@rancher-server:/tmp/cattle-debug.log /tmp/cattle-debug.log

With a local copy of the logs you can either examine the logs using your favorite text editor or you can upload them elsewhere for examination.

Conclusion

With all of our Rancher server metrics dumping into Sysdig Cloud along with our logs dumping into Logstash it has made it easier for multiple people to quickly view and analyze what was going on with the Rancher servers.  In HA Rancher environments with more than one rancher-server running, it also makes filtering logs based on the server or IP much easier.  Since we use 2 hosts in our HA setup we can now easily filter the logs for only the server that is acting as the master.

As these container based grow up, they also become much more complicated to troubleshoot.  With better logging and monitoring systems in place it is much easier to tell what is going on at a glance and with the addition of the monitoring solution we can be much more proactive about finding issues earlier and mitigating potential problems much faster.

Read More

Docker for Mac file system performance summary

One of the more controversial topics right now in the Docker community is the issue surrounding file system performance in the Docker for Mac application.

For a very long time users have been forced to use workarounds to speed up performance when dealing with slow read and write times.  For example, this thread has been open on the Docker forums for over a year now, describing the problem and various workarounds users have found during that time.  There have been blog posts describing various optimizations, as well as scripts and tools to alleviate some of the frustration around slow file system performance on Docker for Mac.

There is a great explanation from the Docker team that lays out the details of the file system performance issues and what the crux of the problem is right now.

At the highest level, there are two dimensions to file system performance: throughput (read/write IO) and latency (roundtrip time). In a traditional file system on a modern SSD, applications can generally expect throughput of a few GB/s. With large sequential IO operations, osxfs can achieve throughput of around 250 MB/s which, while not native speed, will not be the bottleneck for most applications which perform acceptably on HDDs.

The article later goes on to highlight the plan to improve performance along with a number of specific items for accomplishing this.

Under development, we have:

  1. A Linux kernel patch to reduce data path latency by 2/7 copies and 2/5 context switches
  2. Increased OS X integration to reduce the latency between the hypervisor and the file system server
  3. A server-side directory read cache to speed up traversal of large directories
  4. User-facing file system tracing capabilities so that you can send us recordings of slow workloads for analysis
  5. A growing performance test suite of real world use cases (more on this below in What you can do)
  6. Experimental support for using Linux’s inode, writeback, and page caches
  7. End-user controls to configure the coherence of subsets of cross-OS bind mounts without exposing all of the underlying complexity

Additionally, with the latest release of the Docker for Mac 17.04-ce-mac7 (April 6 2017) client, a new :cached flag has been introduced for volume mounts to help with read times for lots of files.  There is also work going on to introduce another :delegated flag to help speed up write times.

Initial user testing of the :cached flag has been good, and shown up to a 4x improvement in some cases.  You can follow this issue on Github to get the most up to date information.  There is some really good detail and discussion going on over there (towards the bottom of the issue is where the new flags are discussed).

Overall I think Docker has done a great job of keeping users informed and updated on the various aspects of the problem and has been steadily making progress in addressing the situation.  The container ecosystem is still very young so there will be growing pains along the way and I think the way that Docker has been handling things has been more than reasonable as they have consistently been making progress on addressing the issue and have been transparent in recent months about what’s going on and how they’re working on the problem.

Read More

Curl on Windows using a Docker wrapper

Does the Windows built-in version of “curl” confuse or intimidate you?  Maybe you come from a Linux or Unix background, and yearn for some of your favorite go-to tools?  Newer versions of Powershell include a cmdlet for interacting with the web called Invoke-WebRequest, which is useful, but is not a great drop in replacement for those with experience in non Windows environments.  The Powershell cmdlets are a move in the right direction to unifying CLI experiences but there are still many folks that have become attached to curl over the years, including myself.  It is worth noting that a Windows compatible version of curl has existed for a long time, however it has always been a nuisance dealing with the zip file, just as using SSH has always been a hassle on Windows.  It has always been possible to use the *nix equivalent tools, it is just clunky.

I found a low effort solution for adding curl to my Windows CLI flow, that acts as a nice middle ground between learning Invoke-WebRequest and installing curl binaries directly, which I’d like to share.  This alias trick is a simple way to use curl for working with API’s and other various web testing in Windows environments without getting tangled in managing versions, and dealing with vulnerabilities.  Just download the latest Docker image to update curl to the newest version, and don’t worry about its implementation across different systems.

Prerequisites are light.  First, make sure to have the Docker for Windows app installed (stable or beta are both fine) as well as a semi-recent version of Powershell.

Next step.  If you haven’t set up a Powershell profile, there are also lots of links and resources about how to do it.   I even wrote about it recently, so I am skipping that step as well.  Start by adding the following snippet to your Powershell profile (by default located in C:\Users\<user>\Documents\WindowsPowerShell\Microsoft.PowerShell_profile.ps1) and saving.

# Curl alias using docker
function Docker-Curl {
   docker run --rm byrnedo/alpine-curl $args
}

# Aliases
New-Alias dcurl Docker-Curl

Then source you terminal and run the curl command that was just created.

dcurl -h

One issue you might notice from the snippet above is that the Docker image is not an “official” image.  If this bothers you (security concerns, etc.), it is really easy to create your own, secure image.  There are lots of examples of how to create minimal images with Curl pre-installed.  Just be aware that your custom image will need to be maintained and occasionally rebuilt/published to guard against future vulnerabilities.  For brevity, I have skipped this process, but here’s an example of creating a custom image.

Optional

To update curl, just run the docker pull command.

docker pull apline-curl

Now you have the best of both worlds.  The built-in Invoke-WebRequest cmdlet provided by Powershell is available, as well as the venerable curl command.

My number one case for using curl in a container is that it has been in existence for such a long time (less bugs and edge cases) and it can be used for nearly any web related task.  It is also much handier to use curl for those with a background using *nix systems, rather than digging around in unfamiliar Powershell docs for similar functionality.  Having the ability to run some of my favorite tools in an easy, reproducible way on Windows has been a refreshing experience while sliding back into the Windows world.

Read More

Generate Certbot certificates with a container

This is a little bit of a follow up post to the origin post about generating certs with the DNS challenge.  I decided to create a little container that can be used to generate a certificate based on the newly renamed dehyrdated script with the extras to make DNS provisioning easy.

A few things have changed in the evolution of Let’s Encrypt and its tooling since the last post was written.  First, some of the tools have been renamed so I’ll just try to clear up some of the names if there is any confusion.  The official Let’s Encrypt client has been renamed to Certbot.  The shell script used to provision the certificates has been renamed as well.  What used to be called letsencrypt.sh has been renamed to dehydrated.

The Docker image can be found here.  The image is essentially the dehydrated script with a few other dependencies to make the DNS challenge work, including Ruby, a ruby script DNS hook and a few Gems that the script relies on.

The following is an example of how to run the script:

docker run -it --rm \
    -v $(pwd):/dehydrated \
    -e AWS_ACCESS_KEY_ID="XXX" \
    -e AWS_SECRET_ACCESS_KEY="XXX" \
    jmreicha/dehydrated-dns --cron --domain test.example.com --hook ./route53.rb --challenge dns-01

Just replace test.example.com with the desired domain.  Make sure that you have the DNS zone added to route53 and make sure the AWS credentials used have the appropriate permissions to read and write records on route53 zone.

The command is essentially the same as the command in the original post but is a lot more convenient to run now because you can specify where on your local system you want to dump the generated certificates to and you can also easily specify/update the AWS credentials.

I’d like to quickly explain the decision to containerize this process.  Obviously the dehydrated tool has been designed and written to be a standalone tool but in order to generate certificates using the DNS challenge requires a few extra tidbits to be added.  Cooking all of the requirements into a container makes the setup portable so it can be easily automated on different environments and flexible so that it can be run in a variety of setups, with different domain names and AWS credentials.  With the container approach, the certs could potentially be dropped out on to a Windows machine running Docker for Windows if desired, for example.

tl;dr This setup may be overkill for some, but it has worked out well for my purposes.  Feel free to give it a try if you want to test out creating Certbot certs with the deyhrdated tool in a container.

Read More

Running containers on Windows

There has been a lot of work lately that has gone into bringing Docker containers to the Windows platform.  Docker has been working closely with Microsoft to bring containers to Windows and just announced the availability of Docker on Windows at the latest ignite conference.   So, in this post we will go from 0 to your first Windows container.

This post covers some details about how to get up and running via the Docker app and also manually with some basic Powershell commands.  If you just want things to work as quickly as possible I would suggest the Docker app method, otherwise if you are interested in learning what is happening behind the scenes, you should try the Powershell method.

The prerequisites are basically Windows 10 Anniversary and its required components; which consist of the Docker app if you want to configure it through its GUI or the Windows container feature, and Hyper-V if you want to configure your environment manually.

Configure via Docker app

This is by far the easier of the two methods.  This recent blog post has very good instructions and installation steps which I will step through in this post, adding a few pieces of info that helped me out when going through the installation and configuration process.

After you install the Win 10 Anniversary update, go grab the latest beta version of the Docker Engine, via the Docker for Windows project.  NOTE: THIS METHOD WILL NOT WORK IF YOU DON’T USE BETA 26 OR LATER.  To check, open your Docker app version by clicking on the tray icon and clicking “About Docker” and make sure it says -beta26 or higher.

about docker

After you go through the installation process, you should be able to run Docker containers.  You should also now have access to other Docker tools, including docker-comopse and docker-machine.  To test that things are working run the following command.

docker run hello-world

If the run command worked you are most of the way there.  By default, the Docker engine will be configured to use the Linux based VM to drive its containers.  If you run “docker version” you can see that your Docker server (daemon) is using Linux.

docker version

In order to get things working via Windows, select the option “Switch to Windows containers” in the Docker tray icon.

switch to windows containers

Now run “docker version” again and check what Server architecture is being used.

docker version

As you can see, your system should now be configured to use Windows containers.  Now you can try pulling a Windows based container.

docker pull microsoft/nanoserver

If the pull worked, you are are all set.  There’s a lot going on behind the scenes that the Docker app abstracts but if you want to try enabling Windows support yourself manually, see the instructions below.

Configure with Powershell

If you want to try out Windows native containers without the latest Docker beta check out this guide.  The basic steps are to:

  • Enable the Windows container feature
  • Enable the Hyper-V feature
  • Install Docker client and server

To enable the Windows container feature from the CLI, run the following command from and elevated (admin) Powershell prompt.

Enable-WindowsOptionalFeature -Online -FeatureName containers -All

To enable the Hyper-V feature from the CLI, run the following command from the same elevated prompt.

Enable-WindowsOptionalFeature -Online -FeatureName Microsoft-Hyper-V -All

After you enable Hyper-V you will need to reboot your machine. From the command line the command is “Restart-Computer -Force”.

After the reboot, you will need to either install the Docker engine manually, or just use the Docker app.  Since I have already demonstrated the Docker app method above, here we will just install the Docker engine.  It’s also worth mentioning that if you are using the Docker app method or have used it previously, these commands have been run already so the features should be turned on already, simplifying the process.

The following will download the engine.

Invoke-WebRequest "https://master.dockerproject.org/windows/amd64/docker-1.13.0-dev.zip" -OutFile "$env:TEMP\docker-1.13.0-dev.zip" -UseBasicParsing

Expand the zip into the Program Files path.

Expand-Archive -Path "$env:TEMP\docker-1.13.0-dev.zip" -DestinationPath $env:ProgramFiles

Add the Docker engine to the path.

[Environment]::SetEnvironmentVariable("Path", $env:Path + ";C:\Program Files\Docker", [EnvironmentVariableTarget]::Machine)

Set up Docker to be run as a service.

dockerd --register-service

Finally, start the service.

Start-Service Docker

Then you can try pulling your docker image, as above.

docker pull microsoft/nanoserver

There are some drawback to this method, especially in a dev based environment.

The Powershell method involves a lot of manual effort, especially on a local machine where you just want to test things out quickly.  Obviously the install/config process could be scripted out but that solution isn’t idea for most users.  Another drawback is that you have to manually manage which version of Docker is installed, this method does not update the version automatically.  Using a managed app also installs and manages versions of the other Docker productivity tools, like compose and machine, that make interacting with and managing containers a lot easier.

I can see the Powershell installation method being leveraged in a configuration management scenario or where a specific version of Docker should be deployed on a server.  Servers typically don’t need the other tools and should be pinned at specific version numbers to avoid instability issues and to make sure there aren’t other programs that could potentially cause issues.

While the Docker app is still in beta and the Windows container management component of it is still new, I would still definitely recommend it as a solution.  The app is still in beta but I haven’t had any issues with it yet, outside of a few edge cases and it just makes the Docker experience so much smoother, especially for devs and other folks that are new to Docker who don’t want to muck around the system.

Check out the Docker for Windows forums if you run into any issues.

Read More