Comments on: Oracle IOPS and HBA Queue Depth http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/ Jeremy Schneider Wed, 08 Sep 2010 23:15:02 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: Gerry Bragg http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-1406 Gerry Bragg Thu, 13 Nov 2008 00:19:52 +0000 http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-1406 How did you like the 3Par? We purchased one last year and it is quite an amazing unit, particularly if you don't want to be managing individual disks. Blew away our EMC Clarion. How did you like the 3Par? We purchased one last year and it is quite an amazing unit, particularly if you don’t want to be managing individual disks. Blew away our EMC Clarion.

]]> By: Craig http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-1346 Craig Thu, 11 Sep 2008 21:29:45 +0000 http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-1346 ever try setting to 16? Sometimes not queing and sending to fast SAN disks speeds things up. When you have a good number of small LUNs assigned as ASM, a lower Que depth can work better. Craig ever try setting to 16? Sometimes not queing and sending to fast SAN disks speeds things up. When you have a good number of small LUNs assigned as ASM, a lower Que depth can work better.

Craig

]]> By: lishan cheng http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-1334 lishan cheng Wed, 13 Aug 2008 19:40:09 +0000 http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-1334 Hi Jeremy Have you tried test with different number of LUNs? Thanks -- LSC Hi Jeremy

Have you tried test with different number of LUNs?

Thanks

–
LSC

]]> By: Jeremy Schneider http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-859 Jeremy Schneider Fri, 11 Apr 2008 18:49:47 +0000 http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-859 Hey Jeremy. My name is also Jeremy Schneider. I'm out of Oshkosh, Wisconsin. I know it sounds like a hoax, but it's not. I'm very into technology and the computers as well. Great Site. Hey Jeremy. My name is also Jeremy Schneider. I’m out of Oshkosh, Wisconsin. I know it sounds like a hoax, but it’s not. I’m very into technology and the computers as well. Great Site.

]]> By: Krishna Manoharan http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-855 Krishna Manoharan Wed, 09 Apr 2008 04:09:35 +0000 http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-855 Hi Jeremy, While higher queue depths do give good results, it can also result in overruns. It may be a better to set a per target limit along with the LUN limits so what a target never gets throttled beyond its capabilities. On the other hand, how do you actually measure the active queue depth (peak and average)? Running iostat -x will give this information, however you could use SWAT to collate it and you could profile your database storage needs pretty well. Thanks Krishna Manoharan http://dsstos.blogspot.com Hi Jeremy,

While higher queue depths do give good results, it can also result in overruns. It may be a better to set a per target limit along with the LUN limits so what a target never gets throttled beyond its capabilities. On the other hand, how do you actually measure the active queue depth (peak and average)? Running iostat -x will give this information, however you could use SWAT to collate it and you could profile your database storage needs pretty well.

Thanks
Krishna Manoharan
http://dsstos.blogspot.com

]]> By: Andralg http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-845 Andralg Tue, 25 Mar 2008 19:44:40 +0000 http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-845 nice work, man nice work, man

]]> By: Alex Gorbachev http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-842 Alex Gorbachev Thu, 20 Mar 2008 03:27:33 +0000 http://www.ardentperf.com/2008/03/13/oracle-iops-and-hba-queue-depth/#comment-842 Hi Jeremy, Sorry for coming back late to this but I've been quite busy recently. We are just about to get our hands dirty on this new cluster so I don't know how it's doing yet but I presume everything is OK - nobody complained so far so you must have done a good job. Interesting results on queue depth. Thanks for sharing that. <i>... while I was waiting for the OS guys to resolve some networking issues... The first spec had an InfiniBand interconnect but....</i> Now imagine how long it could be to wait if that would be not simple Gigabit Ethernet but, indeed, Infiniband. ;-) Cheers, Alex Hi Jeremy,

Sorry for coming back late to this but I’ve been quite busy recently.
We are just about to get our hands dirty on this new cluster so I don’t know how it’s doing yet but I presume everything is OK – nobody complained so far so you must have done a good job.
Interesting results on queue depth. Thanks for sharing that.

… while I was waiting for the OS guys to resolve some networking issues… The first spec had an InfiniBand interconnect but….

Now imagine how long it could be to wait if that would be not simple Gigabit Ethernet but, indeed, Infiniband. ;-)

Cheers,
Alex

]]>