WL#3087: Cluster Disk Data Space Used

Status: Un-Assigned   —   Priority: Medium

We need a tool that tells us how much of a data file we have consumed. If I
insert a 1,000,000 into a database, I should be able to run a command and have
it tell me the percentage used and free in the data file.

Running out of room in the data file will bring cluster processing to a halt
until expanded. This is not HA. DBA/Admins should be able to find out a head of
time that they are running out of room so they can expand by adding another file
and not incur down time.
Currently INFORMATION_SCHEMA.FILES show #free extents
   These are extents that can be used by any table.

It is not applicable to show anything other on a per file basis.
I think this example shows why.
Suppose you have a datafile with only 2 extents, each of size 1Gb
One extent is allocated to table T1 and is full.
The other extent is allocated to table T2 and is empty.

If one present this a file 50% full (or empty) it is very missleading
  as one can not insert into _any_ other table than T2.

  I.e even in 1Gb is "empty" I can't make any inserts...
  I think this alternative is out of the question.


Since this is not possible, one must extend information show per table.
I.e. find a way of informating that table T1 has 0 free bytes in the extents
  that it currently has allocated and the table T2 can make 1Gb of inserts
  before it need to allocate a new extent.

The two alternatives that I can think of now is information_schema.tables
  or "show table info" (or alike).
I have not however examined either of these, 
  But I suggest that the person that gets assigned to this WL should do that.


Also note that I think that what's currently in INFORMATION_SCHEMA.FILES 
  is enough to satisfy the second paragraph in the describtion above.

But as discussion has proved not the intent of the entire WL.