Elasticsearch is unhealthy after uploading vcf

[GeorginaAmbriz]

Hello,
I made a local install of Seqr and I succeed in uploading the reference data and a vcf file with "Option #2: annotate and load on-prem". This means I made all points of local install and upload instructions from 1 to 5. I came back today to checkup on seqr installation, for this I made
docker compose ps
and I saw that elastic search was "unhealthy". I updated seqr with:

docker-compose pull
docker-compose up -d seqr

But I had this error message:
dependency failed to start: container elasticsearch is unhealthy

seqr didn't restart and I had this output:

docker compose up -d seqr
[+] Running 3/3
 ✔ Container geo-redis-1     Healthy                                                                                                                                    0.0s 
 ✔ Container geo-postgres-1  Healthy                                                                                                                                    0.0s 
 ✘ Container elasticsearch   Error                                                                                                                                      0.0s 
dependency failed to start: container elasticsearch is unhealthy

Please find attached the log of elasticsearch that I got with:
docker compose logs -f elasticsearch > elastic_search_logs.txt

elastic_search_logs.txt

I really appreciate any help you can provide about this issue,

Georgina Jimenez Ambriz
Bioinformatics Core
Neuro

[hanars]

try starting the services sequenctially and see if that fxes the issue:

docker compose down seqr
docker compose up elasticsearch
docker compose up seqr

[GeorginaAmbriz]

Hello,
Thank you for your answer. I ran the commands that you gave me, this is the output:

docker compose down seqr
[+] Running 2/0
 ✔ Container geo-seqr-1  Removed                                                                    0.0s 
 ! Network geo_default   Resource is still in use   

docker compose up elasticsearch
[+] Running 1/0
 ✔ Container elasticsearch  Running                                                                 0.0s 
Attaching to elasticsearch

The last command has been running for tow hour and it's still running. Is it normal? Can it take a lot of time?

[hanars]

what about curl -s elasticsearch:9200/status

[GeorginaAmbriz]

nothing happened, I didn't have an output this time neither.

[hanars]

Okay, can you try fully restarting seqr and then running docker-compose ps and send me the output

[GeorginaAmbriz]

Hi,
When I try to fully restart seqr I have this output:

docker compose up -d seqr
[+] Running 3/3
 ✔ Container geo-redis-1     Healthy                                                                                                                         0.0s 
 ✘ Container elasticsearch   Error                                                                                                                           0.0s 
 ✔ Container geo-postgres-1  Healthy                                                                                                                         0.0s 
dependency failed to start: container elasticsearch is unhealthy

And this is the output of docker compose ps:

NAME             IMAGE                                                  COMMAND                                                        SERVICE         CREATED        STATUS                    PORTS
elasticsearch    docker.elastic.co/elasticsearch/elasticsearch:7.16.3   "/bin/tini -- /usr/local/bin/docker-entrypoint.sh eswrapper"   elasticsearch   26 hours ago   Up 26 hours (unhealthy)   0.0.0.0:9200->9200/tcp, :::9200->9200/tcp, 9300/tcp
geo-postgres-1   postgres:12.6                                          "docker-entrypoint.sh postgres -c listen_addresses=*"          postgres        4 days ago     Up 4 days (healthy)       5432/tcp
geo-redis-1      redis:6.2.7                                            "docker-entrypoint.sh redis-server"                            redis           4 days ago     Up 4 days (healthy)       6379/tcp

[hanars]

so a full restart should not have your elasticsearch running for 26 hours and redis up for 4 days. Please fully take down all the components and restart from scratch