Google Spanner reflections

Google Spanner (warning: PDF) has made some noise recently, and I got around to giving their paper a quick read. I was happy to discover that it describes a solution to a problem I've been thinking about for awhile. I was happy to read that my musing were heading in the right direction. The key (as the Googlers are at pains to point out, and I was also considering) is how to handle distributed co-ordination of commits (in a scalable manner).

Spanner introduces the concept of tracking time by a combination of instants, intervals, and assertions about the two - on the understanding that there's uncertainty of the "current" time. As I read through the paper, I tend to think that there's no particular need for atomic and GPS clocks (although they're definitely desirable); so long as the confidence in the current time can be established. The effect of rising uncertainty in the current time would be a reduction in performance, I imagine...

It turns out that the venerable NTP protocol is able to provide measures of uncertainty; so one could re-implement Spanner without atomic clocks (at least, initially), and could then assault the various questions of electing "leaders" and manging reads and writes in a scalable manner.

Apache Zookeeper is the first thought for the leader election, but they talk about clusters of 4 machines being a large deployment, and about "within this data centre", etc. Let's reserve Zookeeper for a use case that fits better; something without Spanner's global ambitions. Spanner's architecture means that this is a key real concern. A closer reading of the paper is necessary to figure out what's involved here; and how it intimately it's related to uncertain timing.

The distributed reads and writes problem does look a bit like Ceph...

Spanner may not be so intractable, after all.

VMware player 4.0.3 on Ubuntu 12.04

VMware player has some well-known issues with setting up the network for player under Ubuntu 12.04 (that is, where Ubuntu is the host OS).

With one small change, I followed the great instructions here: http://www.kartook.com/2012/05/vmware-virtual-network-device-unable-to-loadcompile-vm-player-4-0-2-in-ubuntu-12-04/

The change is that the script you download has the recognised version numbers hardcoded in the header.

Before you run the script (patch-modules_3.2.0.sh), open the file using your preferred text editor, and change the 4.0.2 to 4.0.3, so it looks like:

   plreqver=4.0.3

The script (and VMware player) seems to run fine for VMware Player 4.0.3 on Ubuntu 12.04. If it should make a difference, I'm running x86_64 Ubuntu 12.04.

Empty an Oracle schema, leave an empty schema

What it says on the tin. Sufficient code analysis will reveal that some bizarre schema dependencies may not be destroyed by this; but experience will show that's not actually a problem :-)

The code lives as a gist over on github. A snapshot is below, for simplicity:

set echo off
set verify off
set serveroutput on size 100000

-- Hosted at http://lastinfinitetentacle.blogspot.com/2012/03/empty-oracle-schema-leave-empty-schema.html

-- Disable all contraints
BEGIN
  FOR c IN
  (SELECT c.owner, c.table_name, c.constraint_name
   FROM user_constraints c, user_tables t
   WHERE c.table_name = t.table_name
   AND c.status = 'ENABLED'
   ORDER BY c.constraint_type DESC)
  LOOP
    dbms_utility.exec_ddl_statement('alter table ' || c.owner || '.' || c.table_name || ' disable constraint ' || c.constraint_name);
  END LOOP;
END;
/

-- remove all objects
declare

cursor dropObjectsCusor is
select 'drop ' || object_type || ' ' || object_name as sqlDropStmt
from user_objects
where object_type <> 'TABLE' and object_type <> 'INDEX'
order by object_type;

cursor dropTablesCusor is
select 'truncate table ' || object_name as sqlTruncTbl,
'drop table ' || object_name || ' cascade constraints' as sqlDropTbl
from user_objects
where object_type = 'TABLE'
order by object_type;

begin

for ob in dropTablesCusor
loop
begin
  execute immediate ob.sqlTruncTbl;
exception when others then dbms_output.put_line('Could not truncate a table.');
end;
begin
  execute immediate ob.sqlDropTbl;
exception when others then dbms_output.put_line('Could not drop a table.');
end;
end loop;

for ob in dropObjectsCusor
loop
begin
  execute immediate ob.sqlDropStmt;
exception when others then dbms_output.put_line('Could not drop some object.');
end;
end loop;


end;
/
purge recyclebin;