RE: Dell-Oracle-Linux: Anyone else run this...because its not working for us!

  • From: brian.x.wisniewski@xxxxxxxxxxxx
  • To: cmarquez@xxxxxxxxxxxxxxxx
  • Date: Fri, 9 Dec 2005 14:42:35 -0500

I've been watching this thread with interest. 

At my last company we had Dell 6650's and 2850's running on various 
clusters, RHAS 3.0, RAC & OCFS.  At times we had lockups of the last node 
in the cluster which is the node on which we run RMAN.  System locks up 
and they have to physically power cycle the machine.  Sysadmin wants to 
blame RMAN, vendors want to blame each other and the problem just 
continues.  I was thinking maybe it had something to do with the ext 
filesystem the backups were getting written to but maybe it's not. 

The disks are all on EMC arrays so the 'caching' will mask any performance 
issues at the disk layer because of the RAID algorithm they chose.  The 
EMC vendor was pretty surprised at the throughput we were able to achieve 
when performing backups - he hadn't seen that kind of performance on Intel 
machines before.  I wonder if the Dell hardware and controllers are part 
of the problem now.

We had these lockups occur in several production and test regions.  All 
Dell hardware, all Red Hat Linux, all RAC.  Other machines running Dell 
and Linux - no problems, Oracle/RAC and RMAN added to the mix doesn't seem 
to play well.

Unbreakable!!!

- Brian





"Marquez, Chris" <cmarquez@xxxxxxxxxxxxxxxx>
Sent by: oracle-l-bounce@xxxxxxxxxxxxx
12/09/2005 01:26 PM
Please respond to cmarquez
 
        To:     <Brandon.Allen@xxxxxxxxxxx>, <tomday2@xxxxxxxxx>, 
<oracle-l@xxxxxxxxxxxxx>
        cc: 
        Subject:        RE: Dell-Oracle-Linux: Anyone else run 
this...because its not working for us!



Allen,

>> ·    Array Subsystem:
>>      o       PERC 3/DC, Firmware 196T

What is your storage array?  We have a PV 220s...although I believe the 
additional PERC*/DC (PCI?) cards we added to the servers are the real 
culprit.
If a PV220 do you run is "split" or "cluster" mode?

>>Maybe there is just a problem with Linux
>>or RAC running on this hardware?

I would like to believe this and I know for sure that our RAC config using 
OCFS seemed much "harder" on the disks.

However;

1.)
I have 2 db system almost identical hardware (PERC4/DC & OCFS vs. PERC3/DC 
& EXT3) one RAC and the other not and we see the SAME ERRORS.

2.)
Again, when this same hardware was on SuSE7-Oracle8, no RAID, PV220 in 
"split mode" we did not have issues.
We run RH3-Oracle9i, RAID5/1, PV220 in "cluster mode" now...so more than 
just the OS changed with this config.

3.)
One of my BKP servers RH3-Oracle9i is using the PERC3/DC cards to the 
PV220s with no RAID for about two weeks not without issue...then again 
this db is a standby and not hit nearly as hard

My point being if it is Linux at fault why would the problem disappear 
when we run no RAID config.

If I have not said it before what we get is MegaRaid module errors at the 
OS level but we believe the module is cooking on bad hardware (in RAID 
use).

???

I don't know, abut what Dell concluded says a lot to me...that is why I 
shared it.

Chris Marquez
Oracle DBA


-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx on behalf of Allen, Brandon
Sent: Fri 12/9/2005 12:10 PM
To: tomday2@xxxxxxxxx; oracle-l@xxxxxxxxxxxxx
Subject: RE: Dell-Oracle-Linux: Anyone else run this...because its not 
working for us!

I've been running Oracle on the following config with no problems for a 
couple years, averaging about 500 IOPS (350 reads/s + 40 writes/sec times 
4 due to the RAID 5 parity = 350+(40*4)=510 IOPS).  The write performance 
is terrible due to RAID5, but other than that it seems to run great. Maybe 
there is just a problem with Linux or RAC running on this hardware?

·       Dell 6650
·       Win2K3 Enterprise Edition
·       4 Xeon 2Ghz CPUs w/ Hyper-Threading
·       3.5 GB RAM
·       Array Subsystem:
        o       PERC 3/DC, Firmware 196T
        o       8 x 73gB drives (10k RPM Ultra320 SCSI)
                o       Virtual Disk 0 - 2 disks, RAID1 - 68.24GB capacity
                o       Virtual Disk 1 - 2 disks, RAID1 - 68.24GB capacity
                o       Virtual Disk 2 - 4 disks, RAID5 - 204.73GB 
capacity
·       Oracle 9.2.0.6 Enterprise Edition





-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx]On Behalf Of Thomas Day
Sent: Friday, December 09, 2005 9:42 AM
To: oracle-l@xxxxxxxxxxxxx
Subject: Re: Dell-Oracle-Linux: Anyone else run this...because its not
working for us!


Dell replaced our controller.  The Dell technician opined as how the
PERC controller probably wasn't up to the load imposed by Oracle.
This was in a RAID 0 configuration.

I guess that's what you get when you go with the lowest bidder.

Looking forward to honing my recovery techniques to a fine edge --- again.
--
//www.freelists.org/webpage/oracle-l



Privileged/Confidential Information may be contained in this message or 
attachments hereto. Please advise immediately if you or your employer do 
not consent to Internet email for messages of this kind. Opinions, 
conclusions and other information in this message that do not relate to 
the official business of this company shall be understood as neither given 
nor endorsed by it.

--
//www.freelists.org/webpage/oracle-l



Other related posts: