[lugm.org] Oracle db on debian 6 ia64

Wed Jun 1 15:06:09 UTC 2011

Oracle is probably a C app. It should work 90% on any
POSIX system.

Also, it makes commercial sense for oracle to work on Linux. If you 
have a client on Linux, with mysql and other services. It's going to 
be more difficult to sell oracle, if it works only on Solaris.

However, I doubt they test it on anything but RedHat or SuSe. 

--- On Wed, 1/6/11, Keshwarsingh Nadan <kn at debian.mu> wrote:

From: Keshwarsingh Nadan <kn at debian.mu>
Subject: Re: [lugm.org] Oracle db on debian 6 ia64
To: "'LUGM Discuss Mailing List'" <discuss at lugm.org>
Date: Wednesday, 1 June, 2011, 9:59

Aha there is nothing wrong is testing! I’m at least proud for having the opportunity to test oracle ee on debian! The issue I had was hardware related (write cache on physical disks), the blade is performing really well , and I highly believe that redhat & suse ent should be dropped.!   Your question regarding centos: I could have used centos, oracle runs great on centos, you need to install X in order to run the oracle installer, installing X on centos means a stupid service called “Bluetooth” will be installed along  with loads of bloatwares else it will fail the dependencies... Removing these services afterwards would cause issues..  Anyways I’m backup by 3 more blades running hp ux 11 on three different  chassis, if ever anything goes wrong with the actual debian blade.  The databases are physically located on a storage, even if the debian blade is down for whatsoever reason, I don’t really care.. I should back to business within
 minutes if you understand what I’m trying to mean. Mounting oracle dbs isn’t difficult at all.  J  From: discuss-bounces at lugm.org [mailto:discuss-bounces at lugm.org] On Behalf Of ashvin saminathen
Sent: Wednesday, June 01, 2011 12:33 PM
To: LUGM Discuss Mailing List
Subject: Re: [lugm.org] Oracle db on debian 6 ia64  nice that you managed to get back on track!

and i must say you are a very courageous guy!!! running 250k$ worth of oracle ee, on 70k$ worth of integrity but using a free officially  unsupported OS!  this is a sign of utmost courage, or well in some cases blatant stupidity (the boundary between the two is sometimes very blurry)!

anyway i hope that you don't come to face any additional issues further down the way, for it does sound like a shame paying 22% yearly oracle support... and well since its on an unsupported OS not getting any "oracle level" support when your system goes hayward!

On 1 June 2011 08:44, Keshwarsingh Nadan <kn at debian.mu> wrote:Hah I managed to get back in track, I rebooted and enabled “physical drive write cache” though it warned me about the “disaster” that might occur if blade is not ups protected. It seems that the box is functioning normally… A quick answer: the motivation here is that I’m tired with hp support, according to hp patches are released and tested, but when you apply, it always breaks something & the crazy unlimited user licenses… Oracle: we are running oracle ee which indeed is damn expensive, moved from oracle se that was on the old rx2620.  I even tried freebsd, unfortunately the fiber daughtercard was incompatible, else I would have carried on with it. Oracle works fine on oracle though its unsupported, anyway I don’t need any “oracle level” support from oracle. Suse ent & red hat is a waste of money. kn From: ashvin.saminathen at gmail.com
 [mailto:ashvin.saminathen at gmail.com] On Behalf Of ashvin saminathen
Sent: Wednesday, June 01, 2011 3:25 AM
To: kn at debian.mu; LUGM Discuss Mailing ListSubject: Re: [lugm.org] Oracle db on debian 6 ia64 just a quick question, not really relevant to your problem.but what motivated the shift from hp-ux to debian, particularly on integrity?
I deployed some rx6600 a few years back, and i can't really see the logic in moving from HP-UX to linux on integrity... even from a financial point of view, since you are running on a 4 processor machine, you must be on oracle EE DB, pretty sure 16 processor oracle licenses and support are much more of a burden than the hp-ux licenses and support...

coming back to your original question, i only worked with prod oracle on solaris or hp-ux, but i found the following article... maybe it can help you : http://www.idevelopment.info/data/Oracle/DBA_tips/Linux/LINUX_8.shtml#Configuring%20Oracle%20User%20Shell%20Limits
also how's your swap configurations?

p.s
if you get fed up of dealing with HP and paying those crazy prices, gimme a call, I know of a great Japanese company who are very accommodating when it comes to servers :)On 31 May 2011 21:22, <kn at debian.mu> wrote:Yea the prompt appears immediately when oracle is not running! Increasing the wide limits helped a lot, to which value can I extend it? I've never dealt with these kind of issues before.
Sent from my BlackBerry® wireless device-----Original Message-----
From: Ajay R Ramjatan <ajay.ramjatan at gmail.com>
Date: Tue, 31 May 2011 18:17:10
To: <kn at debian.mu>; LUGM Discuss Mailing List<discuss at lugm.org>
Subject: Re: [lugm.org] Oracle db on debian 6 ia64

Silly question perhaps but you did narrow down the culprit being
Oracle right? If Oracle is not running, then the prompt appears
immediately?

On Tue, May 31, 2011 at 18:03,  <kn at debian.mu> wrote:
> Hello Systems Admins / Engineers,
>
> I need some advise;
>
> I've migrated from an old hp integrity rx2620 server running hp ux 11b) to an integrity bl870c i2 blade with the following specs:
>
> 4x itanium 9340
> 704gb ram (44x 16gb ram sticks)
> 4x300 gb @ 15k rpm sas (raid 1 with 2 online spares for os only)
> OS: debian 6.0.2
>
> Oracle 11g r2 is being run on the server and databases stored on an hp eva 4400.
>
> The problem started when the blade celebrated its first day in production! The shell prompt appears 15mins after hitting the password! Tried to log in via ssh, it took approx one hour for the prompt to display!
>
> Oracle advised to dismount all databases and increase file descriptors, I adjusted the oracle user a soft & hard limit of 100000 / 9999999 in /etc/security/limits.conf. Appended fs.file-max = 9999999 in /etc/sysctl.conf , did a sysctl -p & rebooted.
>
> Prompts now appears after approx 8 mins on console, 35 mins via ssh..
>
> To which extent can I increase the wide limits?
>
> Any help would be greatly appreciated.
>
> kn
> Sent from my BlackBerry® wireless device
>__________________________________________________________
> Linux User Group of Mauritius (LUGM) Discuss mailing list
> Website: http://lugm.org
> Mailing list archive: http://lugm.org/pipermail/discuss_lugm.org/
> Forum: http://lugm.org/forum/
> IRC: #linux.mu on Freenode
>
__________________________________________________________
Linux User Group of Mauritius (LUGM) Discuss mailing list
Website: http://lugm.org
Mailing list archive: http://lugm.org/pipermail/discuss_lugm.org/
Forum: http://lugm.org/forum/
IRC: #linux.mu on Freenode 
__________________________________________________________
Linux User Group of Mauritius (LUGM) Discuss mailing list
Website: http://lugm.org
Mailing list archive: http://lugm.org/pipermail/discuss_lugm.org/
Forum: http://lugm.org/forum/
IRC: #linux.mu on Freenode  
-----Inline Attachment Follows-----

__________________________________________________________
Linux User Group of Mauritius (LUGM) Discuss mailing list
Website: http://lugm.org
Mailing list archive: http://lugm.org/pipermail/discuss_lugm.org/
Forum: http://lugm.org/forum/
IRC: #linux.mu on Freenode
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://discuss.lugm.org/pipermail/discuss_discuss.lugm.org/attachments/20110601/103d6303/attachment.html>