Another fun day not in the office. A week on leave to relax on my birthday and I’ve been working harder than ever, at least it’s on my own stuff.
My main backup server was reporting hard drive errors on an external drive. As it’s ntfs I couldn’t fsck it under linux so plugged it into a windows server and let windows chkdsk it. Then detached the device cleanly under windows before unplugging it and moving it back to the linux server.
Bugger, the linux server couldn’t detect it anymore, kept thinking it was a usb hub instead of a disk drive. Moved it back to the windows server, it thought it was new hardware as well, also not a disk drive. Bother.
Moved it back to the linux server, at least that logs messages I can read.
usb cable unplug, plug back in. I detects it as a usb device but now a disk drive.
power cable unplug/replug, no messages at all; but with my hand on top of the disk unit I can feel it spin before stopping again.
Ok, I have a spare disk, identical model/size (don’t we all, those of us that are paranoid about backups anyway). Unpackaged that and took the disk unit through. Plugged in the existing power and usb cables and… same thing. Yippee, the disk has not dies the problem is obvious.
Repackaged my ‘new’ disk back into its box for the day I will need it, along with all its cables as I want to keep all the new gear in one place.
And as I knew the power cord was good, as I could feel the disk spin when that was plugged in; I pulled out a spare nine-pin USB cable I had been carrying about to charge my mp3 player at work (cough, couldn’t have it’s bettery flat driving home could I) and plugged that into the disk and the linux server, and there is was, a SCSI disk drive was detected, available, and mountable. So I don’t need to use my spare disk just yet.
But who would have guessed, moving the USB cable from the usb port in a PC to a long usb cable plugged into the powered hub connected to the windows server, and then moving it back to the linux server again, either broke or fried the usb cable.
But a new USB cable has sorted that, which is good, I didn’t really have to start with a new disk again as I would have to rerun the bare metal recovery backups for all my servers to repopulate the disk; although I will anyway I suppose as it’s been over a month since the last set and I can’t be 100% sure the windows chkdsk left the existing ones intact.
And as an aside, while I was working on that I also sorted out the issue I had with c3270, it seems to be incredibly fussy with whats in the /etc/hosts file. Once I removed a duplicate entry for the local server hostname it just started working; so the MVS38J image is available again on port 3270 for those who have 3270 emulators and what to play in it.
