planb-cassandra

planb-cassandra copied to clipboard

Extend should erase_on_boot

1

Currently when extending an existing ring, all user-data gets copied to new instances, including erase_on_boot set to false. this needs to mounting issues during taupage boot, as the new volumes...

jbspeakr

tune/discuss GC G1 options

1

``` JVM_OPTS="$JVM_OPTS -XX:G1HeapRegionSize=32 -XX:G1ReservePercent=20 -XX:+PrintGC -XX:+PrintGCDateStamps -XX:+UseG1GC -XX:MaxGCPauseMillis=200 -XX:InitiatingHeapOccupancyPercent=55 -XX:ParallelGCThreads=10 -XX:ConcGCThrea ds=7 -XX:G1RSetUpdatingPauseTimePercent=10" ``` switched config also on single node to use offheap_buffer

Jan-M

During the cluster creation, the function create_cluster.launch_instance creates a new EBS volumen per each Cassandra node without checking first if a volume exists with the same name. Then, a volume...

SergioGonzalezSanz

Support multiple rings in one AWS region

3

by possibly setting the snitch dc prefix? http://docs.datastax.com/en/cassandra/2.1/cassandra/architecture/architectureSnitchEC2_t.html

Jan-M

When prompting to update a node, we could show the current compaction tasks so that the operator can decide to either proceed with updating or skip the node and let...

a1exsh

Currently fullstop will create violations of type `APPLICATION_VERSION_DOES_NOT_HAVE_A_VALID_ARTIFACT` if it scans a Cassandra cluster which was updated to use a newer docker image version.

a1exsh

Tag EC2 instances with Cassandra type and environment

2

This would allow us to improve monitoring.

a1exsh

Configuration without new docker image

2

We should enable passing a map of configuration options to the docker image. This will remove the need to build a new docker image (which would require a rolling update)...

slitsche

Auth failure from docker entry point script is confusing

1

Let's try to set the admin password in a different way that doesn't spam the logs on every start with these: ``` Connection error: ('Unable to connect to any servers',...

a1exsh

Currently the create_cluster code selects seed nodes from the lowest private IP addresses, and update_cluster sorts the instances to update by private IP. The overall effect is that when performing...

a1exsh