Not sure about software raid or what you have on the disk nor how long it's been running, but you may want to "df -i" and check inode use. Other than that, I have had problems with hot swap scsi drives not making a good connection. Try pulling it out and reinserting it. Lights/LEDs is the best way to see what's going on at the hardware level of scsi disk. No light on one disk, definately got a problem. jay On Tue, Nov 25, 2003 at 05:03:13PM -0800, list@xxxxxxxxxxxxxxx wrote: > Forrest, > Thank you so much for taking the time to test this in a "safe" manner. It is > definitely possible that a cable came loose during transportation, and if this > does not work then I will be going down to the DC shortly :-) > > I will try this early this coming morning so if the machine crashes hard I can > switch DNS and be up relatively shortly. > > I will let you guys know how this turns out so that others searching the > archives can benefit from my situation. > > Thanks again, > > Peter > > > > > Quoting "Taylor, ForrestX" <forrestx.taylor@xxxxxxxxx>: > > > On Tue, 2003-11-25 at 16:02, list@xxxxxxxxxxxxxxx wrote: > > > Forrest, > > > Thanks for taking the time to do a real simulation of what would happen. I > > have > > > a backup server in place at a different facility that can take over, but > > that > > > would require a DNS change that not all ISP's will obey. > > > > > > This is a very strange situation, and I have learned that before shipping > > off a > > > machine one should always make sure that if a drive is lost everything > > will > > > occur as expected. > > > > These are the results that I got, and they may differ from what you get, > > because I cannot really simulate exactly what you have (I am not sure > > why sda is not part of the array). > > > > I created a RAID 1 array for / and /home. I then turned off the machine > > and unplugged one of the disks. When I turned it back on, it booted up > > properly, and I saw something similar in /proc/mdstat that you had--only > > one disk show up. In /var/log/messages, I get a message at boot time > > that says: > > > > md0: former device sda1 is unavailable, removing from array! > > > > I created a RAID partition and I added it to the array: > > > > raidhotadd /dev/md0 /dev/sdb5 > > > > Then I can cat /proc/mdstat and see md0 is rebuilding (recovery = %) > > > > It may have been possible that sda was knocked loose during > > shipping/mounting, and it wasn't working at boot time. Someone may have > > noticed this, and reseated the drive. This would explain why sda is not > > a part of the array, but you can see it with fdisk. > > > > Anyway, in another terminal, do `tail -f /var/log/messages` to see the > > kernel messages, and try: > > > > raidhotadd /dev/md0 /dev/sda1 > > raidhotadd /dev/md1 /dev/sda2 > > > > Check /proc/mdstat to verify that they are rebuilding. If so, you > > should be good to go. > > > > Forrest ---end quoted text--- -- Shrike-list mailing list Shrike-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/shrike-list
