> On 10/1/07, jepoy < [EMAIL PROTECTED]> wrote: > hi guys, > > i just want to ask this technology if its LACP or virtual ip thing. Like if i > have 2 > servers running the same service and i want to have a redundancy on these > service.The 2 servers(active and backup) have different physical ip address > but > the service is running on a virtual ip address, wherein if one of them is > down the > service will just switch to the active server. Is this applicable to linux? > anyone > tried these before?
hi jepoy, yes it is applicable and depends on the service you are running to achieve your goal... i will try to avoid the politics going on here with your thread and look at the bright side to help you :-> LACP or link aggregation control protocol... is a protocol used by IEEE 802.3ad for bonding several nics on a host into one virtual nic or channel...this provides load balancing, fail over and increase bandwidth for that host only. Therefore LACP is not appropriate for your needs here. Virtual IP... unlike with LACP doing on a single host, VIP is load balancing, fail over and increase bandwidth on multiple hosts or servers... although you can combine VIP and LACP... but it is not advisable... LVS is doing a VIP but you need another box for the front end of your real servers or server farm... one LVS is a point of failure and you go back to your basic question again of redundancy... of course you need two LVS to achieve this... there are lots of solutions for this and i said it depends on the service you are running and these are the solutions that i can think of... 1. Common Address Redundancy Protocol (CARP)... carp is an improve version of VRRP (virtual router redundancy protocol) that came from HSRP (hot standby router protocol)... carp came from the bsd and linux copies the idea also... the trick there is that... for example you have two servers named SERVER A and SERVER B... and you need two ip addresses named IP 1 and IP 2... you instruct carp on SERVER A that IP 1 is the active and IP 2 is the standby while on SERVER B that IP 1 is the standby and IP 2 is the active... if any of the server fails, the active IP of that failed server became the active IP (previously on standy IP mode) of the surviving server... thus IP 1 and IP 2 are still available... with this you dont need a front-end LVS for highly availability... see www.ucarp.org for linux version... 2. Single System Image (SSI)... ssi is a single virtual operating system across multiple servers... you just run one app and let the ssi do the highly availability for you across the servers where HA and LVS were integrated on this project.... see http://openssi.org/cgi-bin/view?page=openssi.html for more info... fooler. _________________________________________________ Philippine Linux Users' Group (PLUG) Mailing List [email protected] (#PLUG @ irc.free.net.ph) Read the Guidelines: http://linux.org.ph/lists Searchable Archives: http://archives.free.net.ph
