Zero Disk space

We are using Elastic for a year now, and we're very pleased about it.
We have 2 nodes sitting on different servers with estimated size of 250GB of data sitting one index with 6 shards each.

every now and then we will run out of disk space and I would delete some doc's and run the optimize with only_expunge_deletes=true command.
but now we reached a point were that command needs addition disk space to make the lucene index marge.
So it failed on one of our nodes :confused:, but the good news is that we do succeeded in the other node and right now our stat is Yellow.

My question is :
Because the failed node has no disk space can I manually delete shards folders, hopping that the working node will reallocate the missing shard (which was already space reduced) back ?
If not which files can I delete in order to give the miserable node enough space to recover and complete the optimize command ?

Thanks in advance.

Yes but it's dangerous.

Ideally you should add more disk, or more nodes, or alter your indexing structure to let you delete old data.

What could happen if I delete for example the folder "0" under the ..\nodes\indices{indexname}\ ?

You lose that shard from that node.

But I will get it's replica from the other node right ?

Assuming there is one, yes.