How We Leveraged DRBD 9 Autopromote for a Simplified Pacemaker Cluster Layout
Historically, NetEye Clusters were configured with DRBD as Master/Slave resources. This led to the following rather cumbersome resource configuration for an N-node cluster:
$SERVICE_drbd_master ( x 1 )
$SERVICE_drbd_master_clone ( x N )
$SERVICE_drbd_fs
$SERVICE_virt_ip
$SERVICE
Note: $SERVICE serves as a placeholder for any Cluster Service running in NetEye 4.
At least ten constraints were necessary to model the required relations (on promote once, collocation 5 times, and after 4 times):
DRBD 9 introduced a nice little convenience feature called “autopromote”. This automatically promotes a node to primary as soon as the DRBD resources’ block device is mounted. If dual primary mode is configured, then this works on up to 2 nodes, otherwise the mount just fails.
We can rely on this behavior to dispose of the ugly Master/Slave part of the above Cluster configuration. The resources can then be reduced to the following:
$SERVICE_drbd_fs
$SERVICE_virt_ip
$SERVICE
The constraints also become a lot simpler:
Why didn’t we just use Resource Groups?
Well, you can’t put Master/Slave resources in a Resource Group. However, as you may have noticed, we just got rid of all master/slave resources. This means we actually can get rid of all constraints! Resource groups already enforce the necessary collocation and order constraints on their members.
The final result is the following 4 Resources ( 1 Group + 3 Members ) per Service/DRBD Device:
$SERVICE_group
$SERVICE_drbd_fs
$SERVICE_virt_ip
$SERVICE
No constraints whatsoever are necessary, and furthermore DRBD is no longer a cluster-controlled service, but can be run and monitored as a local service on each individual node.
Hi, my name is Benjamin, and I'm Software Architect in the Research & Development Team of the "IT System & Service Management Solutions" Business Unit of Würth Phoenix.
I discovered my passion for Computers and Technology when I was 7 and got my first PC. Just using computers and playing games was never enough for me, so just a few months later, started learning Visual Basic and entered the world of Software Development. Since then, my passion is keeping up with the short-lived, fast-paced, ever-evolving IT world and exploring new technologies, eventually trying to put them to good use. I'm a strong advocate for writing maintainable software, and lately I'm investing most of my free time in the exploration of the emerging Rust programming language.
Author
Benjamin Gröber
Hi, my name is Benjamin, and I'm Software Architect in the Research & Development Team of the "IT System & Service Management Solutions" Business Unit of Würth Phoenix.
I discovered my passion for Computers and Technology when I was 7 and got my first PC. Just using computers and playing games was never enough for me, so just a few months later, started learning Visual Basic and entered the world of Software Development. Since then, my passion is keeping up with the short-lived, fast-paced, ever-evolving IT world and exploring new technologies, eventually trying to put them to good use. I'm a strong advocate for writing maintainable software, and lately I'm investing most of my free time in the exploration of the emerging Rust programming language.
Running Ollama locally or on dedicated hardware is straightforward until you need to know whether a model is actually loaded in RAM, how fast it generates tokens under load, or when memory consumption reaches a threshold that affects other workloads. Read More
Hi everyone! Today I'd like to share with you an investigation we undertook related to ingesting Open Telemetry data in Elasticsearch, while maintaining tenant segregation from start to end. The Scenario Let's imagine we have multiple customers, where in this Read More
SNMP monitoring is the standard method for obtaining information and metrics from network devices. Typically, we focus on extracting data from a single interface to monitor its status, traffic, or errors. But in many cases, we’re only interested in getting Read More
In the ITOA module we fixed a bug that prevented the Performance Graphs to be shown in the Monitoring host and service page. List of updated packages grafana, grafana-autosetup, grafana-configurator and grafana-neteye-config to version 12.4.1_neteye3.29.2-1
Creating a GitHub organization is easy. Creating a public one that is actually well-structured, secure, and maintainable over time… not so much. At the beginning, it feels like a simple task: create the org, push some repositories, maybe define a Read More