Description
So far, we've been using only block storage (AWS EBS), which can't be accessed in ReadWriteMany mode.
Because of that, we need to have one repository cache for each worker - see #210
On MP+ we can also use AWS EFS file storage (basically NFS - network file system) which can be accessed in ReadWriteMany mode.
Hence, after we have both, stg & prod on MP+, we can consolidate all the RWO worker volumes (1+1 per long-running worker) we use for the repository cache into one(?) RWX.
And we can change workers from StatefulSet to Deployment at the same time, I guess.
Benefit
The long-running workers will use significantly less disk space (1 volume instead of 4 for 2 l-r workers)
Importance
gain/low, impact/low
Workaround
Participation
Description
So far, we've been using only block storage (AWS EBS), which can't be accessed in
ReadWriteManymode.Because of that, we need to have one repository cache for each worker - see #210
On MP+ we can also use AWS EFS file storage (basically NFS - network file system) which can be accessed in
ReadWriteManymode.Hence, after we have both,
stg&prodon MP+, we can consolidate all the RWO worker volumes (1+1 per long-running worker) we use for the repository cache into one(?) RWX.And we can change workers from
StatefulSettoDeploymentat the same time, I guess.Benefit
The long-running workers will use significantly less disk space (1 volume instead of 4 for 2 l-r workers)
Importance
gain/low,impact/lowWorkaround
Participation