This is getting easier. Logstash is running and so is the Prometheus exporter for it. Kibana, on the other hand, is a nightmare. I can’t get it do anything other than spew security warnings. I made them go away – by adding “basic security” to Elasticsearch and Kibana – and it broke all the logstashes.Continue reading “Logstash Up”
Category Archives: Tech
Gone Public
It’s not quite ready, but at the request of an interested party, I made the github repository public. One more (successful) rebuild-from-scratch and I’m calling it good. In addition to getting Pulsar Manager running, finally, I wrote (trivial) tests for both Pulsar and Cassandra. They work! BLEK is still missing. Jaeger is a possibility. But,Continue reading “Gone Public”
Rebuilt!
A few bobbles, but I rebuilt the whole thing from instructions. About three pages of instruction-update notes, which I will deal with later. It’s bed time, again. [WordPress did want larger featured images]
Cassandra Up, Again
A three node, three rack, one data center Cassandra cluster is running! It has a few configuration values that require IP addresses, yuck! All the cluster software is up! What’s left? Install the Docker monitor for Prometheus (cAdvisor). Install BLEK. Cross-reference build scripts against written instructions. Throw it all away and rebuild from scratch. MakeContinue reading “Cassandra Up, Again”
Grafana Up
Docker really isn’t all that easy. If you want all defaults, it’s dandy. If you want customize anything (e.g. save log files or configure it), then it’s not. Grafana has some sort of strange user/permission thing going on. Sorted. The pretty picture is the Stack OS metrics dashboard (gathered by node_exporter, collected by Prometheus, andContinue reading “Grafana Up”
Prometheus Up, Again
Prometheus was mostly working on the ops machine. I tried to Docker it. Prometheus, itself, came up, but none of the cluster.test names resolved. Getting github working proved its worth: git push on the ops machine, then git pull on another, and voila! all the files I need are there. Docker actually proved worthwhile here,Continue reading “Prometheus Up, Again”
On github
No, you cannot see it, yet – it’s not done. I got it rearranged, unified and up there – and made git work on all the VMs, so I can edit from anywhere.
Pulsar Up, Again
While baking cookies, I rewrote the broker scripts for the DNS changes. On the bright side, I never updated them to cluster.dev, so it was straight from IPs to cluster.test. I was confused about why the Bookies had port 6650 mapped when it could not be configured. Turns out, that is a Broker port thatContinue reading “Pulsar Up, Again”
Up, Again
And cluster.test is up and running as far as BookKeeper. All six http:// URLs work. All scripts and documents now use cluster.test instead of cluster.dev. (Google is evil!) There were a few bobbles along the way, but I’m getting reasonably good at this, unfortunately. Discovery: When you send –restart always to a docker run command,Continue reading “Up, Again”
Google Is EVIL
I cannot connect a web browser to my new entirely internal domain because I named it cluster.dev and Google has registered “.dev” and hard-coded into browsers that it must be secure. I’m not screwing with TLS infrastructure. Now I need to rename the whole thing. “.test” is supposed to be reserved. We’ll see. Many thanksContinue reading “Google Is EVIL”