HPE EVA Storage

Re: Secure Path 3.0F in Active-Active failover time

 
David_450
Occasional Advisor

Secure Path 3.0F in Active-Active failover time

Looking for some input regarding the fail over of Secure path 3.0F with EVA 4000-6000
I have ran some test and found out that it takes about 27 seconds.
With Secure path 3.0E active-passive it was instant need some help with this.
I have created the VG01 with all 4 links.


Thank you
David S
5 REPLIES 5
Mike Roger
Advisor

Re: Secure Path 3.0F in Active-Active failover time

Hi,

Have you tried Secure Path 3.0F SP1 it does address the Active-Active failover

following is the link:

http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareIndex.jsp?lang=en&cc=us&prodNameId=1825336&prodTypeId=12169&prodSeriesId=471465&swLang=8&taskId=135&swEnvOID=7

GL
Mike
Mohanasundaram_1
Honored Contributor

Re: Secure Path 3.0F in Active-Active failover time

Hi David,

I may be totally wrong. But this 27 seconds seems to be the expected behaviour as the default pvtimeout value is 30 seconds for the LVM PVs in HP-UX.

WIth regards,
Mohan.
Attitude, Not aptitude, determines your altitude
David_450
Occasional Advisor

Re: Secure Path 3.0F in Active-Active failover time

From what i can see it is the behaviour of PVlinks but having Secure Path 3.0F sp1 I would expect to have imidiate fail over.
and I am not


any other information would appriciated

Thank you
Mohanasundaram_1
Honored Contributor

Re: Secure Path 3.0F in Active-Active failover time

HI David,

I should explained my answer a bit more. The 3.0E version is meant for active/passive controllers. Here the pvlinks are not used.

But the EVA4000-6000 requires Securepath version which can handle active/active controllers. Therefore 3.0F SP1 is required. Interestingly, when you install the active/active component of the securepath, it is called as Autopath.

This autopath does not mask your device paths like the 3.0E versions. Therefore all the possible device paths are displayed in ioscan. You have to include all the alternate paths as pvlinks and then Autopath can provide load balancing.

Because of this feature, I was mentioning about the pvtimeout value.

My guess is, 3.0E version was handling path failover at Securepath level. 3.0F SP1 version gives this path failover to be handled by pvlinks. I hope somebody can validate this guess.

With regards,
Mohan.
Attitude, Not aptitude, determines your altitude
David_450
Occasional Advisor

Re: Secure Path 3.0F in Active-Active failover time

Well yes Autopath is the current use in Secure Path 3.0F but that only gives load balancing no instant fail over still I think it uses PV links 27 seconds.

What is the use of 2 HBA in cluster if this is the case

Thank you for the help