I logged in this morning, only to see my main server was rebooted last night..i was quite pissed at first, as i keep telling the other admins never to reboot or shutdown the system. But I investigated more and saw that none of them were logged in at the time, and that none of them initated a reboot. I have this in my utmp/wtmp log:
me ttyp0 max48-ppp227.evt Sun May 23 11:00 - 11:44 (00:44) reboot system boot 2.0.36 Sun May 23 04:32 (07:25) queenj ttyp0 max48.evt0.firet Sat May 22 23:28 - 02:59 (03:30) reboot system boot 2.0.36 Sat May 22 23:02 (05:17) 'me' is an admin ...the time seems to be a bit off, it was offset by an hour when i signed on, so i ran ntpdate to update it. anyways..i figure that hostname '2.0.36' was the kernel rebooting the system? I do have watchdog support enabled, and have watchdog running. Nobody would be at the office at 4am to reboot the server, im sure, although i won't know until they reply to my emails(im 1500 miles away from the machine) I did see a flood of these errors in my kernel log file: May 23 03:08:00 galactica kernel: Unable to load interpreter May 23 03:10:00 galactica last message repeated 2 times May 23 03:14:00 galactica kernel: Unable to load interpreter May 23 03:15:00 galactica kernel: Unable to load interpreter shows those errors right before both reboots..anyone know what they could mean? I'm runnin linux 2.0.36+securelinux SMP on a dual processor P2-266. also, is it..a good idea (and how would i do it) to make the system run ckraid automatically when it starts up, i noticed in the bootuplog that all my raid arrays are not clean(i suppose because the kernel issued a hard reboot preventing the arrays from being stopped) REGISTER_DEV sda5 to md0 done REGISTER_DEV sdb5 to md0 done md: 09:00: raid array is not clean -- run ckraid REGISTER_DEV sda6 to md1 done REGISTER_DEV sdb6 to md1 done md: 09:01: raid array is not clean -- run ckraid REGISTER_DEV sda7 to md2 done REGISTER_DEV sdb7 to md2 done md: 09:02: raid array is not clean -- run ckraid REGISTER_DEV sda8 to md3 done REGISTER_DEV sdb8 to md3 done md: 09:03: raid array is not clean -- run ckraid REGISTER_DEV sda5 to md0 done REGISTER_DEV sdb5 to md0 done I'm running raid 1(mirroring) accross 2 drives(Ultra 2 SCSI). This is the first time this has happened since the server came online..almost 2 months ago. Nothing in the startup(kernel wise, etc) has been changed. I figure maybe we were under a process table attack or something, but without better logs i cant tell. im glad the box came back up..phew. and if it was watchdog..very impressed that it actually worked. nate
