Known issues on GCP

Specific Kubernetes version required for ElasticBLAST 1.0.0 users

On April 19, 2023 GKE upgraded the default version of Kubernetes, so ElasticBLAST version 1.0.0 users on GCP will have to add the following configuration parameter to their ElasticBLAST configuration file(s):

[cloud-provider]
gke-version = 1.24

This will not be necessary for ElasticBLAST versions greater than 1.0.0. This configuration will not work after October 31 2023 as Kubernetes version 1.24 will reach its end-of-life. Please, update ElasticBLAST to the latest version.

Upgrade required for GCP ElasticBLAST users

ElasticBLAST versions prior to 1.0.0 will stop working because version 1.21 of kubernetes at GKE has reached end of life on January 31, 2023:

https://cloud.google.com/kubernetes-engine/docs/release-schedule

To ensure ElasticBLAST continues to work for you on GCP, please upgrade ElasticBLAST to its latest version.

Files left in cloud storage

ElasticBLAST uses cloud storage to temporarily store query sequences and internal logs and metadata so that they are easily accessible during its operation. Sometimes deleting these files after the search is not successful. To double check and delete them, please run the commands below. ELB_RESULTS below represents the location where your results are stored.

gsutil ls gs://${ELB_RESULTS}/query_batches  # to list query files
gsutil -m rm gs://${ELB_RESULTS}/query_batches/*  # to delete query files

gsutil ls gs://${ELB_RESULTS}/logs  # to list log files
gsutil -m rm gs://${ELB_RESULTS}/logs/*  # to delete log files

gsutil ls gs://${ELB_RESULTS}/metadata  # list metadata files
gsutil -m rm gs://${ELB_RESULTS}/logs/*  # to delete metadata files

Too many query batches leads to failed execution

ElasticBLAST divides query sequences into batches and searches them in parallel. The Batch length parameter controls the size of a single batch. If the ElasticBLAST configuration leads to more than 5,000 query batches, ElasticBLAST will exit with an error message prompting you to increase Batch length parameter to decrease the number of query batches.