Despite the title, this is actually a technical post about Oracle, disk I/O and Exadata & Oracle In-Memory Database Option performance. Read on :)
If a car dealer tells you that this fancy new car on display goes 10 times (or 100 or 1000) faster than any of your previous ones, then either the salesman is lying or this new car is doing something radically different from all the old ones. You don’t just get orders of magnitude performance improvements by making small changes.
Perhaps the car bends space around it instead of moving – or perhaps it has a jet engine built on it (like the one below :-) :
I have been teaching the Enkitec Exadata Administration Class this week and made an interesting observation I thought was worth sharing with regards to IO Resource Management on Exadata.
I have created a Database Resource Manager (DBRM) Plan that specifically puts a resource consumer group to a disadvantage. Actually, quite severely so but the following shouldn’t be a realistic example in the first place: I wanted to prove a point. Hang-on I hear you say: you created a DBRM plan-the post has IORM in the subject though: what gives? Please allow me to explain.
Exadata offers 3 different ways to implement IORM to the keen engineer:
(the details are investigated and specific to Oracle’s database implementation on Linux x86_64)
Exadata IO: This event is not used with Exadata storage, ‘cell single block physical read’ is used instead.
Despite p3 listing the number of blocks, I haven’t seen a db file sequential read event that read more than one block ever. Of course this could change in a newer release.
Just a quick post on a new Exadata feature called Zone Maps. They’re similar to storage indexes on Exadata, but with more control (you can define the columns and how the data is refreshed for example). People have complained for years that storage indexes provided no control mechanisms, but now we have a way to exert our God given rights as DBA’s to control yet another aspect of the database. Here’s a link to the 126.96.36.199 documentation which resides in the Data Warehousing Guide: Zone Map Documentation
Zone Maps are restricted to Exadata storage by the way (well probably they work on ZFS and Pillar too). Have a look at the Oracle error messages file:
The Flash Cache Mode still defaults to Write-Through on Exadata X-4 because most customers are better suited that way – not because Write-Back is buggy or unreliable. Chances are that Write-Back is not required, so we just save Flash capacity that way. So when you see this
Last week I’ve gotten a question on how storage indexes (SI) behave when the table for which the SI is holding data is changed. Based on logical reasoning, it can be two things: the SI is invalidated because the data it’s holding is changed, or the SI is updated to reflect the change. Think about this for yourself, and pick a choice. I would love to hear if you did choose the correct one.
First let’s do a step back and lay some groundwork first. The tests done in this blogpost are done on an actual Exadata (V2 hardware), with Oracle version 188.8.131.52.6 (meaning bundle patch 6). The Exadata “cellos” (Cell O/S) version is 184.108.40.206.1.140529.1 on both the compute nodes and the storage nodes.
I know, posts about up-coming user group meetings are not exactly exciting, but it’s good to be reminded. You can’t beat a bit of free training, can you?
On Monday 14th I am doing a lightning talk at the 4th Oracle Midlands event. The main reason to come along is to see Jonathan Lewis talk about designing efficient SQL and then he will also do a 10 minute session on Breaking Exadata (to achieve that aim I suggest you just follow the advice of the Oracle Sales teams, that will break Exadata for you pretty efficiently!).
Anyone who has looked at Exadata might ask the question, and I did so too. After all, cell smart table scan is in wait class User IO so there should be more, right? This is what you find for a smart scan:
NAME PARAMETER1 PARAMETER2 PARAMETER3 WAIT_CLASS ------------------------------ -------------------- -------------------- ------------------------------ --------------- cell smart table scan cellhash# User I/O cell smart index scan cellhash# User I/O
Compare this to the traditional IO request:
I recently was involved in an investigation on a slow-running report on an Exadata system. This was rather interesting, the raw text file with the query was 33kb in size, and SQL Developer formatted the query to > 1000 lines. There were lots of interesting constructs in the query and the optimiser did its best to make sense of the inline views and various joins.
At the Accenture Enkitec Group we have a couple of Exadata racks for Proof of Concepts (PoC), Performance validation, research and experimenting. This means the databases on the racks appear and vanish more than (should be) on an average customer Exadata rack (to be honest most people use a fixed few existing databases rather than creating and removing a database for every test).
Nevertheless we gotten in a situation where the /etc/oratab file was not in sync with the databases registered in the cluster registry. This situation can happen for a number reasons. For example, if you clone a database (RMAN duplicate), you end up with a cloned database (I sincerely hope), but this database needs to be manually registered in the cluster registry. This is the same with creating a standby database (for which one of the most used methods is to use the clone procedure with a couple of changes).