This is a discussion on Redundant Interconnects on 10g RAC within the Oracle Database forums, part of the Database Server Software category; --> Hi All, I've been setting up a simple 2 node RAC on sun hardware using Solaris 9 and CRS, ...
| |||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| Hi All, I've been setting up a simple 2 node RAC on sun hardware using Solaris 9 and CRS, using 2 interfaces on each node as interconnects. Each interconnect is connected to it's partner node via a different gig switch. The issue I'm seeing when testing failover is loss of a single interconnect results in the second node rebooting, it's as though the RAC isn't aware of the other remaining active interconnect network. Is this normal behaviour?, or a likely configuration fubar on my part. I've trawled google but can't find much detail on this to be honest. I'd rather not go down the IPMP route on the interconnects as it would mean trunking the 2 switches I have at present. Any information you might be able to share on this would ge great! Cheers, Matt |
| |||
| On Apr 19, 10:15 am, Matt <matt.b...@ntlworld.com> wrote: > Hi All, > > I've been setting up a simple 2 node RAC on sun hardware using Solaris > 9 and CRS, using 2 interfaces on each node as interconnects. > > Each interconnect is connected to it's partner node via a different > gig switch. > > The issue I'm seeing when testing failover is loss of a single > interconnect results in the second node rebooting, it's as though the > RAC isn't aware of the other remaining active interconnect network. Is > this normal behaviour?, or a likely configuration fubar on my part. > I've trawled google but can't find much detail on this to be honest. > > I'd rather not go down the IPMP route on the interconnects as it would > mean trunking the 2 switches I have at present. > > Any information you might be able to share on this would ge great! > > Cheers, > Matt Hi matt, i'm not sure, but i think that oracle doesn't manage redundant private interconnect. The best thing to do i think is to let Operatating system manage it with "channel bonding". I'm not system administrator, but i think that this is right way to do solve your problem Bye from Italy Cristian Cudizio http://cristiancudizio.wordpress.com http://oracledb.wordpress.com |
| |||
| Matt, If you want RAC to use both the network cards, you may want to set cluster_interconnects parameter. However be warned that this is just for load balancing purposes and not for fail overs. If you are looking for a fail over at network card level, you may have to use appropriate OS bonding/teaming techniques and in your case IPMP is the way to go. -- Best Regards, K Gopalakrishnan Co-Author: Oracle Wait Interface, Oracle Press 2004 http://www.amazon.com/exec/obidos/tg.../-/007222729X/ Author: Oracle Database 10g RAC Handbook, Oracle Press 2006 http://www.amazon.com/gp/product/007146509X/ |
| |||
| On Apr 19, 8:17 pm, K Gopalakrishnan <kaygo...@gmail.com> wrote: > Matt, > > If you want RAC to use both the network cards, you may want to set > cluster_interconnects parameter. However be warned that this is just > for load balancing purposes and not for fail overs. If you are > looking for a fail over at network card level, you may have to use > appropriate OS bonding/teaming techniques and in your case IPMP is the > way to go. > > -- > Best Regards, > K Gopalakrishnan > Co-Author: Oracle Wait Interface, Oracle Press 2004http://www.amazon.com/exec/obidos/tg/detail/-/007222729X/ > > Author: Oracle Database 10g RAC Handbook, Oracle Press 2006http://www.amazon.com/gp/product/007146509X/ I feared this may be the case, thanks very much for the clarification! cheers, Matt |
| ||||
| Matt wrote: > Hi All, > > I've been setting up a simple 2 node RAC on sun hardware using Solaris > 9 and CRS, using 2 interfaces on each node as interconnects. > > Each interconnect is connected to it's partner node via a different > gig switch. > > The issue I'm seeing when testing failover is loss of a single > interconnect results in the second node rebooting, it's as though the > RAC isn't aware of the other remaining active interconnect network. Is > this normal behaviour?, or a likely configuration fubar on my part. > I've trawled google but can't find much detail on this to be honest. > > I'd rather not go down the IPMP route on the interconnects as it would > mean trunking the 2 switches I have at present. > > Any information you might be able to share on this would ge great! > > Cheers, > Matt You must do operating system level NIC bonding. -- Daniel A. Morgan University of Washington damorgan@x.washington.edu (replace x with u to respond) Puget Sound Oracle Users Group www.psoug.org |