Couchbase with Docker Compose: Unable to insert - DurabilityImpossibleError - docker

Question

I'm trying to set up Couchbase as part of a collection of servers using Docker-Compose. The sole purpose of this is for local application development.
The problem is that, once set up, I'm unable to write to the database. Insert and Upsert operations give me a DurabilityImpossibleError.
Docker compose file:
version: '3.4'
services:
...
couchbase-db:
image: couchbase/server
volumes:
- ./docker-data/couchbase/node3:/opt/couchbase/var
- ./provision/couchbase:/opt/startup/
ports:
- 8091:8091
- 8092:8092
- 8093:8093
- 8094:8094
- 11210:11210
The startup bash script, run after building, is an attempt to perform database setup without requiring a manual step:
#!/bin/bash
# Enables job control
set -m
# Enables error propagation
set -e
# Run the server and send it to the background
/entrypoint.sh couchbase-server &
# Check if couchbase server is up
check_db() {
curl --silent http://${1}:8091/pools > /dev/null
echo $?
}
# Variable used in echo
i=1
# Echo with
log() {
echo "[$i] [$(date +"%T")] $#"
i=`expr $i + 1`
}
# Wait until main server is ready
until [[ $(check_db 127.0.0.1) = 0 ]]; do
>&2 log "Waiting for Couchbase Server to be available ..."
sleep 1
done
couchbase-cli cluster-init -c localhost:8091 \
--cluster-username Administrator --cluster-password password \
--cluster-password password --services data,index,query --cluster-ramsize 512 \
--cluster-index-ramsize 256 || true
couchbase-cli setting-cluster -c localhost:8091 -u Administrator -p password \
--cluster-username Administrator --cluster-password password \
--cluster-password password --cluster-ramsize 512 \
--cluster-index-ramsize 256;
couchbase-cli setting-cluster -c localhost:8091 \
-u Administrator -p password --cluster-username Administrator \
--cluster-password password --cluster-ramsize 512 --cluster-index-ramsize 256;
curl -v POST http://localhost:8091/pools/default/buckets \
-u Administrator:password \
-d name=organisations \
-d bucketType=couchbase \
-d ramQuotaMB=512 \
-d durabilityMinLevel=majorityAndPersistActive
curl -v -X POST -u Administrator:password \
http://localhost:8091/settings/indexes \
-d indexerThreads=4 \
-d logLevel=verbose \
-d maxRollbackPoints=10 \
-d storageMode=plasma \
-d memorySnapshotInterval=150 \
-d stableSnapshotInterval=40000
# Need to wait until query service is ready to process N1QL queries
echo "$(date +"%T") Waiting ........."
sleep 20
# Create bucket1 indexes
echo "$(date +"%T") Create bucket1 indexes ........."
cbq -u Administrator -p password -s "CREATE PRIMARY INDEX idx_primary ON \`organisations\`;"
cbq -u Administrator -p password -s "CREATE INDEX idx_type ON \`organisations\`(_type);"
If I try to add a document via the web interface, I get:
Errors from server: ["Unexpected server error, request logged."]
If I try to add a document via the JavaScript SDK, I get:
DurabilityImpossibleError durability impossible
details:
{
name: 'DurabilityImpossibleError',
cause: LibcouchbaseError: libcouchbase error 308
at Object.translateCppError
(/app/node_modules/couchbase/dist/bindingutilities.js:174:21)
at /app/node_modules/couchbase/dist/connection.js:245:54 {
code: 308
},
context: KeyValueErrorContext {
status_code: 0,
opaque: 0,
cas: CbCas {
'0': <Buffer 00 00 00 00 00 00 00 00>
},
key: '22738bd4-7972-4370-85a3-71399d96ef05',
bucket: '',
collection: '',
scope: '',
context: '',
ref: ''
}
}
I've also attempted to send the following settings with the insert/upsert, to no effect:
insertOptions: {
durabilityLevel: 0,
durabilityPersistTo: 1,
durabilityReplicateTo: 0,
timeout: 5000,
},
My most recent attempt at a fix was to build a cluster of 3 nodes within docker-compose, and call the API commands to "add server" as part of a build script. However, "add server" takes a static IP, so the second time I run the servers, the IPs change and the database becomes unresponsive. I do get a functioning database on that first run though.
I'm looking for either a fix for a single node system (or an idea of where I'm going wrong), or a way of getting a cluster working in Docker-Compose after a down/up cycle. Anything that will give me a stable environment to develop in.
Thanks!

Answer 1

The bucket is created with -d durabilityMinLevel=majorityAndPersistActive, By default bucket enables replica with 1.
For single node cluster you will not have enough data nodes to satisfy durability (https://docs.couchbase.com/server/current/learn/data/durability.html). You can disable replica via UI and rebalance to take affect or change the bucket setting not include Minimum durability.
I have no idea about the 3 node docker compose error.