SecurePath/AutoPath Questions

[Michelle Barton]

We are using securepath here, and I have some questions:

1~How do I choose between shortest service time and round robin? Not how to configure it, but rather, how do I determine which will work better in my environment?

2~I added a number of disks on my server, yet they are not showing in the autopath display all even after an autopath discover. They show:
AutoPath: Device /dev/dsk/c26t5d3 is not an AutoPath device
AutoPath: Unable to communicate with the device. Please check the device path.

Interesting FWIW, 500 disks display correctly, it's the remaining 116 that don't.

Thanks!
Michelle

[Hoang Minh Tuan]

Hi Michelle,

I don't know what exactly questions you are, but in my point of experiences, SecurePath is used for Compaq storage products like EVA3000, 5000 while AutoPath is used for HP storag products like VA7410..
1. Using service time or round robin is depended on your system requirements. Service time is for the best I/O access while round robin is for the best availability (failover/failback and load balancing)
2. You may restart AutoPath daemon or add device to Autopath device manually.

HMT

[Cem Tugrul]

From one of my prr serv which uses EVA5000
disks;
baan01:/#spmgr display
Server: baan01 Report Created: Fri, Jun 17 09:30:21 2005
Command: spmgr display
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Storage: 5000-1FE1-5000-F2D0
Load Balance: On Auto-restore: On Balance Policy: Round Robin
Path Verify: On Verify Interval: 30
HBAs: td0 td1 td2 td3

baan01:/# swlist|egrep -i secure
CPQswsp A.3.0E.00F.00F HP StorageWorks Secure Path Device Driver and utilities for HSx class of Disk Arrays

As you seen the version of secure path installed on this server...

You can change service time&round robin
parameters from ;


/sbin/spmgr set -a on|off (auto restore) |
-b on|off (load balancing) |
-b rr|li|ls|lb (load balancing algorithm)
(rr - Round Robin,
li - Least I/O,

Good Luck,

[John Payne_2]

I had our DBA team spent 2 weeks running tests under the different load balancing policies. They did identical tests, measured the results, and drew their conclusion for our environment. (1 was clearly better for what we were doing.)

Really, unless you try, you won't be sure what's best for you. (I don't know what you are doing with your SAN...)

Hope it helps

John