Two options for Zenoss 5 installation are available. First one is manual installation. It’s recommended if you want to know more about Zenoss 5. If you haven’t used Docker container concept before, then you may find it too complex or boring when you are installing n-th Zenoss. This was also the reason why I’ve created a second option – Zenoss 5 auto-deploy script.
Zenoss 5 auto-deploy script provides an easy option for installation of the free Zenoss 5 version (Zenoss 5 Core); however, the commercial version (Zenoss 5 Resource Manager) is supported as well.
Typical cases where I recommend Zenoss 5 auto deploy script:
- You want to test Zenoss 5 only, and you don’t need to know what is under the hood.
- You are comfortable with the Zenoss 5 installation, so another installation is only wasting your time.
- You are developing Zenpacks, templates, etc., so Zenoss 5 installation/operation is not important for you.
You probably haven’t noticed that Zenoss 5 auto-deploy script has also implemented telemetry which reports installation issues and successes. If you have noticed then I’m happy, because you are one of many who has read the provided README file. Let’s go to check what we can mine from our statistics.
Average user needs seven attempts for successful installation. It sounds terrible, but it’ll be obvious when you see what the typical user is doing. Average installation takes 41 minutes. Keep in mind that it depends mainly on your internet connection speed. Zenoss Core 5.0.8 download Docker images (Zenoss Core, HBase, OpenTSDB, etc.), which have total virtual size of 5.7GB. One piece of advice: Be patient.
Real example of successful Zenoss 5 Core installation in full HD resolution:
Don’t do it
Zenoss auto-deploy script uses colored highlights. Only one color is important to you, the red one. If you have some red information, then you know that it is a disaster. Maybe you don’t have enough memory, CPUs, or something similar. Fortunately, the script is not very strict and will tell you what the problem is, and it’ll ask you if you want to continue. This option is there only for Zenoss 5 experts who know what they are doing. Many “experts” ignore red warnings and that is the best way to fail.
Typical ignored red warnings:
- At least 20GB of memory is recommended
Blank Zenoss 5 installation (without any monitored devices) consumes 10GB of memory. It means that if you have less than 10GB of memory it won’t start your Zenoss 5 even for testing. Maximum used memory from statistic is 125GB. That’s extreme, and it’s not the best solution on how to scale Zenoss 5. You should scale horizontally (more Zenoss machines in your “cluster”) and not vertically (more powerful machines). Recommended memory is 20GB; it is a perfect start for production monitoring. Many people are blaming Zenoss for such a huge memory requirement. It’s not fair. Zenoss 5 is designated for perfect horizontal scaling, and due to many Docker containers/Zenoss daemons, there are memory duplications. That’s the reason for such a huge memory demand. Hopefully, technology will solve this problem in the future. Keep your eyes on VMWare Photon container virtualization, and their container memory deduplication.
- At least 4 CPUs are recommendedBecause you have container technology in place, you can start many lightweight “virtual machines” at once. It means many processes are starting at once, so your CPU load will be high. I have recorded max CPU load 23 on machine with 8CPUs. It means CPU load 184 for 1CPU machine, and I’ll guess that Zenoss start will be incredible slow in this case. Please use the recommended 4CPUs.
- Filesystem X is required
Some features (Dockers, backups) depends on used filesystem. Follow recommended filesystem, e.g. btrfs for /opt/serviced/var/volumes. Snapshoting really doesn’t work if your filesystem doesn’t support it (xfs, ext4, simfs, …). Remember also for recommended size of your file systems. People have usually problem with underevaluated size of /var/lib/docker, where are stored all Docker images.
- 10 daily successful installations on average
- 70 daily installations attempts on average
- 66% RHEL/CentOS 7, 33% Ubuntu 14
- 5.51CPU average in successful installations (max 64CPUs)
- 15.9GB RAM average in successful installations (max 125GB)
- 41minutes is average duration of a successful installation (max 5.2 hours – probably slow network)
- Number of Zenoss 5 installation attempts per country
- Top 10 world cities in term of number Zenoss 5 installation attempts
Santiago 26(1.58%) Hong Kong 25(1.52%) London 22(1.33%) Bangkok 19(1.15%) Wilmington 19(1.15%) Ho Chi Minh City 19(1.15%) Zurich 18(1.09%) Harbin 18(1.09%) Boston 18(1.09%)
Word at the End
As you can see there is a lot of issues, which can be easily avoided if you follow the recommended settings. I’ve tried to create script which is user friendly (used colors, error messages, workarounds in the script, hints). Unfortunately I can’t cover all problems that can happen during installation. Please use follow outputs from the auto deploy script first (e.g. setup password for ccmaster, DNS/hosts settings). If you still have a problem, then use the community forum/IRC. I’ll do my best, but due to high volume of issues, I have cancelled all personal requests for installation helps. If you need commercial help, then do not hesitate to contact me. I’m open for commercial R&D activities, such as Zenoss deployment with Kubernetes/VMWare Photon/CoreOS, etc.