redo logs and control files

Posted on 2014-03-02
Last Modified: 2014-03-10
can anyone elaborate (in laymans management freindly terms) the risks associated with storing all redo logs and control files on the same drive? I have seen a few articles stating they should be stored on seperate discs, can I ask why? do oracle themselves have a view on this, and any artciles to back their view up?
Question by:pma111
Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

  • Help others & share knowledge
  • Earn cash & points
  • Learn & ask questions
  • 2
  • 2
  • 2
  • +1
LVL 22

Accepted Solution

Steve Wales earned 167 total points
ID: 39899042
I don't know that I have any articles handy but the main reason is fault protection.

If you have redo logs and control files on different devices, and you love one device, you still have a copy since redo logs and control files are redundant copies of themselves.

If you have everything on a single drive, and it fails you lose everything.

If you lose a disk and they are spread out over multiple devices, even if you lose the device with the data files, you still should be able to recover up to the point of failure if you have your redo logs still available somewhere.

Also the control files contain information about your backups if you are doing RMAN and no catalog database so you want to make sure that you have them protected from faults as well.

All on one device is a single point of failure.  Lose it and you lose everything.

Author Comment

ID: 39899048
albeit not a storage guru (nor oracle), does it matter if your server has raided storage? i.e. would that mitigate the need to store them across multiple systems? or is that a dangerous strategy, as if the entire server went boom, same problem occurs... and do you spread them across servers, or across different drives in the same server?

also if you have routine backups, does that again mitigate the impact?  is this more towards minimizing how much data you lose in relation to hardware failure? which if you do daily backups would be a days worth at most, as if the server goes boom I assume everything goes down..
LVL 22

Expert Comment

by:Steve Wales
ID: 39899080
Even on raided storage I tend to move things around onto different devices as much as possible.  What happens if you have a failure in the raid controller ?  Or the enclosure where the disks are spontaneously combusts ?

Admittedly, if you're on a SAN or a VM, there's a whole lot more there that is invisible to the DBA.   If the SAN admin assigns you a couple of disks you have no idea if they are on separate physical devices or the same one.

For backups, you're only as good as the last time the backup was taken.  If you have a backup, and a copy of your redo logs and the backup of your archive log destination you can restore to the point of failure (I had a sysadmin disconnect my datafile drive once in error.  Between the backups and the online redo logs that were still on another volume I recovered right up to the point where the disconnect occurred and the users lost zero data).

It would be a very rare installation where the users would be happy with any data loss, even if it was "only a day".

They might be able to "live with it", but I doubt they'd be happy and you never want to be the DBA who has to tell his boss that you didn't have sufficient redundancy or backups built into your disaster recovery plan to mitigate as much loss as is humanly possible.

Of course there more disaster proof you need your system to be the more it costs.  You certainly want to hope that your bank has multiple redundant control in place including all sorts of hot offsite copying of data by the second as it happens (but I'm getting off track now).
Windows Server 2016: All you need to know

Learn about Hyper-V features that increase functionality and usability of Microsoft Windows Server 2016. Also, throughout this eBook, you’ll find some basic PowerShell examples that will help you leverage the scripts in your environments!

LVL 23

Assisted Solution

David earned 167 total points
ID: 39899094
Hi pma, you're asking good questions.  I think you have the right idea in mind about reducing risk, and some level of risk is acceptable in a business case -- be it the length of unscheduled interruptions, or how much data loss is affordable.  I know of people who work  on a project that requires two DBAs on the primary RAC for redundancy, another two on the COOP RAC in another location, and of course, separate backups on both installations.

So, both the primary and secondary sites have SAN storage, j, with logs and data files striped across virtual drives -- for redundancy to avoid ANY SINGLE POINT OF FAILURE.

Routine backups?  Of course it helps, but when was the last time your management declared a full emergency simulation and you had to fully restore from that backup?  What if you also had to install the O/S and network onto a cold server in another location?  And the primary support person / people weren't available, so that a less-experienced person had to rely upon the available run-time documentation.  Remember, all it takes is a fire or flood somewhere in the building to force the evacuation and quarantine of your data center.

Lastly, all it takes is for one part of the whole process to fail.  I participated last year in a root cause analysis for an organization that took a four-day unplanned outage.  With some effort, the team determined the fault lay with a prior firmware upgrade on the interconnect switch.  The primary site tried to fail over to secondary, and failed, because the remote/backup site's switch had not been included in the firmware change.
LVL 37

Assisted Solution

by:Geert Gruwez
Geert Gruwez earned 166 total points
ID: 39899698
everything can fail, even hardware with a failover system
like a raid with 2 discs going corrupt at the same time
or a core switch not failing over because there's a bug in the failover software

this simple question should be checked at all levels, from hardware to people's skills:
how much do you allow your redundancy to cost and what is the minimum level of redundancy required

Author Comment

ID: 39899821
is it common to seperate the oracle software and all the relevant database files onto different drives as well?
LVL 37

Expert Comment

by:Geert Gruwez
ID: 39900147
dunno about who's doing what in common, but in general i have 5 drives
C: Operating system, D: oracle software, G: control files, redo files, archive logs, H: control files, redo logs, data files, I: backups
whenever wants to backup the server, i indicate they can only backup C, D and I
I tell them i'll do the rest.

backup software with for oracle built specific agents doesn't always work too well

that's just one reason to make a distinction ... failures and redundancy is still the main one

Featured Post

What is SQL Server and how does it work?

The purpose of this paper is to provide you background on SQL Server. It’s your self-study guide for learning fundamentals. It includes both the history of SQL and its technical basics. Concepts and definitions will form the solid foundation of your future DBA expertise.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

When it comes to protecting Oracle Database servers and systems, there are a ton of myths out there. Here are the most common.
Shell script to create broker configuration file using current broker Configuration, solely for purpose of backup on Linux. Script may need to be modified depending on OS-installation. Please deploy and verify the script in a test environment.
This video shows how to set up a shell script to accept a positional parameter when called, pass that to a SQL script, accept the output from the statement back and then manipulate it in the Shell.
This video shows how to copy an entire tablespace from one database to another database using Transportable Tablespace functionality.

749 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question