Hi y'all! I am hoping for a good discussion here. ...I have 2 systems here...different hardware, drives the same size, both systems RAID1 arrays this sort of thing. Both running fully updated fc3 with exactly the same servers running. able to switch between shutting down servers on one turning them on on the other...utilizing an aliased ip address that is constant to the public no matter which server is running. Bob Chiodini has been guiding me thru all the HA techniques that he himself uses. One system is the master, one is the slave...slave keeps pinging the master...if slave doesn't get the answer from the master it wants...then the machines change roles...slave is master and master is slave...that's very simply put but generally how it's done. The big problem here is syncing the data. The master takes in the data. The data has to end up on the slave. almost like a RAID1 array on 2 different systems but it can't be totally like that because of hardware issues. I have thought about keeping track of processes running...decide if it's a hardware related process or not. If it's not hardware then it has to be synced. Keep track of disk activity. Monitor disk activity...what process is writing what? if it's NOT hardware related then it's a go! lock the write when it's finished, could clamscan first, copy to the slave. This is strictly hobby stuff for me at present. I want to just see if I can actually do all this with all stuff on the fedora cds. I don't know if this is possible but I am kinda narrow minded bout fedora, course I can't be open minded or my brains would fall out! I can see rsync used in conjunction w/cfengine. cfengine has a daemon, cfd, will let you pull on demand. I dunno anything bout rsync of cfengine, this just caught my eye. I really would like some ideas if y'all would. thx, John Rose -- Trust Everyone...just make sure you brand your cattle
