Hello, Recently my department had a SAN installed, and I am in the process of setting up one of the first Linux machines connected to it. The machine is running Red Hat Enterprise AS4 (x86_64), which uses Linux kernel version 2.6.9-11.ELsmp. The SAN shows up twice in the kernel, as /dev/sdb and /dev/sdc. /dev/sdb is inaccessible (I get a bunch of "Buffer I/O error on device sdb" kernel errors), but /dev/sdc works fine. According to the administrator of the SAN, the SAN shows up twice because there are two paths to the device, each going through a different storage processor (SP). At any time, only one SP is active, which is why /dev/sdb is inaccessible but /dev/sdc works fine. The administrator informed me that it is possible for the paths to switch (i.e. /dev/sdc becomes inactive and /dev/sdb becomes active), so I need to have some kind of multipathing software installed. He told me to use PowerPath, but I'd rather not have to reinstall or rebuild kernel modules every time there is a kernel upgrade, so I'm looking into Linux's built-in multipath support. I followed the very straightforward instructions available here: http://www.centos.org/docs/4/html/rhel-ig-s390-multi-en-4/s1-s390info-raid.html#S2-S390INFO-MULTIPATH I created /etc/mdadm.conf, then ran: mdadm -C /dev/md0 --level=multipath --raid-devices=2 /dev/sdb /dev/sdc Initially after creation, "mdadm --detail" reports that both paths are up and active: Number Major Minor RaidDevice State 0 8 16 0 active sync /dev/sdb 1 8 32 1 active sync /dev/sdc However, when I actually start using the device, /dev/sdb is removed and is set faulty (which makes sense since /dev/sdb is inaccessible): Number Major Minor RaidDevice State 0 0 0 -1 removed 1 8 32 1 active sync /dev/sdc 2 8 16 -1 faulty /dev/sdb My question is this: If the paths are switched while my machine is up, will Linux be smart enough to check if /dev/sdb has become active when /dev/sdc becomes faulty? Or will the whole device fail? thanks, Jim Faulkner - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
