- Newest
- Most votes
- Most comments
Hello,
You might receive a ClusterBlockException
error for the following reasons:
- Lack of available storage space
- High JVM memory pressure
[+] Troubleshooting Amazon OpenSearch Service - ClusterBlockException - https://docs.aws.amazon.com/opensearch-service/latest/developerguide/handling-errors.html#troubleshooting-cluster-block
Lack of available storage space
If one or more nodes in your cluster has storage space less than the minimum value of 1) 20% of available storage space, or 2) 20 GB of storage space, basic write operations like adding documents and creating indexes can start to fail.
To avoid issues, monitor the FreeStorageSpace
metric in the OpenSearch Service console and create CloudWatch alarms to trigger when FreeStorageSpace
drops below a certain threshold. GET /_cat/allocation?v
also provides a useful summary of shard allocation and disk usage. To resolve issues associated with a lack of storage space, scale your OpenSearch Service domain to use larger instance types, more instances, or more EBS-based storage.
High JVM memory pressure
When the JVMMemoryPressure metric exceeds 92% for 30 minutes, OpenSearch Service triggers a protection mechanism and blocks all write operations to prevent the cluster from reaching red status. When the protection is on, write operations fail with a ClusterBlockException
error, new indexes can't be created, and the IndexCreateBlockException
error is thrown.
When the JVMMemoryPressure metric returns to 88% or lower for five minutes, the protection is disabled, and write operations to the cluster are unblocked.
High JVM memory pressure can be caused by spikes in the number of requests to the cluster, unbalanced shard allocations across nodes, too many shards in a cluster, field data or index mapping explosions, or instance types that can't handle incoming loads. It can also be caused by using aggregations, wildcards, or wide time ranges in queries.
To reduce traffic to the cluster and resolve high JVM memory pressure issues, try one or more of the following:
- Scale the domain so that the maximum heap size per node is 32 GB.
- Reduce the number of shards by deleting old or unused indexes.
- Clear the data cache with the
POST index-name/_cache/clear?fielddata=true
API operation. Note that clearing the cache can disrupt in-progress queries.
In general, to avoid high JVM memory pressure in the future, follow these best practices:
- Avoid aggregating on text fields, or change the mapping type for your indexes to
keyword
. - Optimize search and indexing requests by choosing the correct number of shards.
- Set up Index State Management (ISM) policies to regularly remove unused indexes.
[+] https://repost.aws/knowledge-center/opensearch-high-jvm-memory-pressure
Relevant content
- asked 2 years ago
- AWS OFFICIALUpdated 9 months ago
- AWS OFFICIALUpdated a year ago
- AWS OFFICIALUpdated a year ago
- AWS OFFICIALUpdated a year ago
We have a 10gig instance. There is 7gig free. We should expand that to at least 20gig?
Can we increase the ebs volume size without losing data?
Hello,
We will need to check the Amazon OpenSearch domain to narrow down the exact cause of
ClusterBlockException
- whether it is Lack of available storage space or High JVM memory pressure before recommending any changes.If you continue to experience the issue after performing basic troubleshooting as discussed previously, please reach out to support engineering for further assistance and recommendations.
Hello,
Yes, you can increase EBS volume without losing data. Changing EBS volume will cause a Blue/Green deployment.
As you might be aware, Amazon OpenSearch uses a blue/green deployment process when updating domains, which follows the below process: → AWS will provision a new set of nodes with the desired configuration. → Once these nodes have finished provisioning correctly, data is migrated from your old nodes to the new ones. → Once this migration process is complete, then the old nodes are terminated, leaving you with the new desired configuration.
[+] Making configuration changes in Amazon OpenSearch Service - Changes that usually cause blue/green deployments - https://docs.aws.amazon.com/opensearch-service/latest/developerguide/managedomains-configuration-changes.html#bg
jvm memory pressure is around 90%. Think need to relieve it