RE: Device mapper multipathing

I would begin by looking at the failback and playing around with that a bit. 
Maybe lower it to, ummm, 5 and give that a try. And changing that driver 
timeout value as well.

Rui Amaral


________________________________
From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On 
Behalf Of Zhu,Chao
Sent: Tuesday, January 11, 2011 8:29 PM
To: LS Cheng
Cc: oracle-l
Subject: Re: Device mapper multipathing

Thanks LS/Rui;
I asked our storage guys to take a look at your sample config and do more test;
We used IBM xiv ARRAY for the test in QA EnV and the configuration was 
recommended by IBM;
Below is our configuration; Any recommendation?
In production we will be using HDS AMS2000 or USPV servers; But we are not at 
that stage yet (yes, weird production/qa different SAN during POC);

devices {
    device {
        rr_min_io 15
        path_checker tur
        product "2810XIV"
        no_path_retry fail
        vendor "IBM"
        polling_interval 3
        path_grouping_policy multibus
        path_selector "round-robin 0"
        failback 15
    }
    device {
        rr_min_io 15
        path_checker tur
        product "NEXTRA"
        no_path_retry  fail
        vendor "XIV"
        polling_interval 3
        path_grouping_policy multibus
        path_selector "round-robin 0"
        failback 15
    }
}
defaults {
    user_friendly_names yes
}


On Tue, Jan 11, 2011 at 5:34 PM, LS Cheng 
<exriscer@xxxxxxxxx<mailto:exriscer@xxxxxxxxx>> wrote:
Hi

Can you provide the storage maker and model?

Normally Storage vendors provides failover policies such as

# For EVA A/A arrays
device {
        vendor                  "HP|COMPAQ"
        product                 "HSV1[01]1 
\(C\)COMPAQ|HSV[2][01]0|HSV300|HSV4[05]0"
        path_grouping_policy    group_by_prio
        getuid_callout          "/lib/udev/scsi_id -g -u /dev/%n"
        path_checker            tur
        path_selector           "round-robin 0"
        prio                    alua
        rr_weight               uniform
        failback                immediate
        hardware_handler        "0"
        no_path_retry           18
        rr_min_io               100
}

##      Device attributes for EMC CLARiiON
device {
     vendor                 "DGC "
     product                "*"
     path_grouping_policy   group_by_prio
     getuid_callout         "/sbin/scsi_id -g -u -s /block/%n"
     prio_callout           "/sbin/mpath_prio_emc /dev/%n"
     path_checker           emc_clariion
     path_selector          "round-robin 0"
     features               "1 queue_if_no_path"
     no_path_retry          300
     hardware_handler       "1 emc"
     failback               immediate

}




On Tue, Jan 11, 2011 at 6:58 AM, Zhu,Chao 
<zhuchao@xxxxxxxxx<mailto:zhuchao@xxxxxxxxx>> wrote:
not sure if anyone else has updated this thread since it was posted;
We ran into issue of delayed path failover (when one path dies/disabled, all IO 
hang for 30sec-2 min);
it wasn't that optimal to have that time IO totally freezed;

We were using redhat 5.4 and also tried oracle enterprise kernel;

Would be great to hear more inputs;

Thx



On Wed, Nov 29, 2006 at 7:08 PM, Luca Canali 
<Luca.Canali@xxxxxxx<mailto:Luca.Canali@xxxxxxx>> wrote:
Hi,

Is anybody who has used device mapper for Linux willing to share their
experience?

I have been using Qlogic multipathing so far with the qla2300_conf
module on RHEL3. No problem with it, but I am moving to RHEL4 and the
needed qla2xxx_conf module is missing from their rmps distribution. So
instead I have done a few tests with device mapper multipathing, seems
to work fine so far, but nothing can beat real world experience.

Thanks,
L.
--
http://www.freelists.org/webpage/oracle-l





--
Regards
Zhu Chao






--
Regards
Zhu Chao



NOTICE: Confidential message which may be privileged. Unauthorized 
use/disclosure prohibited. If received in error, please go to www.td.com/legal 
for instructions.
AVIS : Message confidentiel dont le contenu peut être privilégié. 
Utilisation/divulgation interdites sans permission. Si reçu par erreur, prière 
d'aller au www.td.com/francais/avis_juridique pour des instructions.

Other related posts: