| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> Re: Virtual RAC on Solaris E15k
Mark,
First of all, if not already done, download, read and forward to your
management
http://www.miracleas.dk/WritingsFromMogens/YouProbablyDontNeedRACUSVersion.pdf.
Furthermore, http://www.miracleas.dk/WritingsFromMogens/HigherStack.htm
might be of interest. Since Mogens is a director of a consulting company,
please advise your management as much complexity as you can get. Holidays
are immenent, therefore systems stay untouched from human beings giving him
far too little to do. Addinx more complexity to your system might ignite
some compensation for that.
Seriously, Tim already mentioned a good way to achieve db redundancy, instead of service redundancy: Data Guard. I m ight be biased here, because I almost make a living out of installing DG-configurations, but Physical Standby might be of help, and can protect you from human errors as well. As your management appearantly already knows, hrdware stops failing when the management orders it to stop failing. However, DBA's and other human idiots touching the system won't listen to management, and will continue making errors in the most uninconvenient ways. DG can help you with a time lag, postponing the recovery of the redolog information that has been received by the standby database. WHen the timelag is long enough, you might even be able to notice an error before it was applied to the standby. That gives you the opportunity to recover a dropped table from the primary. Redundant storage, including Netapp's snapmirror and EMC's Timefinder will happily forward all your errors synchronously or asynchronously to all storage attached. I've set up tens of DG configurations now, and I think this might help your management really with their problems. However, don't use Logical Standby (LSB) for HA purposes, but Physical Standby. LSB is not stable enough for HA yet. What might help you to convince your management is to come up with a list of all problems you can think of (network outages, SAN/NAS/DAS outages, power outages, etc. etc., and then name the countermeasure to survive that problem with your configuration. That might give some insight in the measures to take in order to survive a disaster. Hardware failures can be skipped, your management just has to order other hardware to stop failing from now on.
Your company has an E15K. That's far from cheap. So there might be some money at stake when something (in your company: personnel) fails. That requires good plans. Data Guard is a good way to recover from human mistakes. I've seen companies not that big, implementing good disater recovery plans, using a redundant machine room in the same building, and a disaster recovery site 40-100 kilometers away.
Regards, Carel-Jan
===
If you think education is expensive, try ignorance. (Derek Bok)
===
-- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------Received on Sat Jun 12 2004 - 18:37:42 CDT
 |
 |