Hello,
Some weeks ago, I had a full disk problem on my dev cluster.
I made some space, reactivated index writing with
PUT /_all/_settings
{
"index.blocks.read_only_allow_delete": null
}
Most of the indexes are doing the rollover correctly since, but today I found that three of them decided to write without the configured alias, and they didn't rollover... leading to a new "no free space available"...
My indexes usually have their name with "-00000" for the rollover alias part.
I don't know what prevents these 3 indexes to be named correctly to rollover...
Any help appreciated