When all Elasticsearch Shards Fail, Go Back to the Basics - ElasticCC

Join Landy Simpson, Software Engineer at Elastic, as she shares her journey of navigating the complexities of Elasticsearch and tackling the cluster block exception as a beginner.

She'll discuss her learnings of core Elasticsearch concepts and how running low on disk space impacts the allocation of shards across nodes. She'll also review her solution and the importance of unlocking indices after freeing up disk space. Lastly, she'll explore the alternate solution of modifying the cluster watermark settings.

Backstory:
Elasticsearch’s ease of use is a double-edged sword. Its out-of-box setup makes it easy for beginners to get started but difficult for beginners to see the need to learn Elasticsearch concepts. It’s not until all shards fail that one realizes it’s important to know what a shard is.

When Landy's boss requested a major feature for a client with a tight deadline, she encountered her first error (the cluster block exception) and panicked. She didn’t know how to make sense of the error or the resources explaining how to solve the error. However, by taking a step back and focusing on understanding what is a cluster, shard, and node, she better understood how to solve the problem.

#Elasticsearch #ClusterBlockException #SoftwareEngineering #TechTips #TechTalk #TechCommunity

[milequinze]

Awesome talk. It was what I was doing: going back to learn some basics. You explained the concepts pretty well, in a way less abstract than most people do.

[morrismutwiri]

Great fundamental concepts, learned a lot. I am now set to run my ELK. Thank you.

[Justicewarrior]

Good job! Thanks for the great explanation

[chinnachinni]

I have faced in production All the Shards gone Down, how long it takes Replica to back(up) in between there is a sync (my question is what is the sync from wr it get data to restore) think 1 of Replica takes as primary, (All the Data exists in Relpica/ else it will get it from somewhere