Server Disk Failure Friday Night
Posted: Sat 24 Sep 2016, 10:53:21If you noticed the site has been down last night and this morning, it was due to a disk failure.
It should be fixed now.
It should be fixed now.
Peak Oil News and Message Boards
Exploring Hydrocarbon Depletion
https://peakoil.com/forums/
Server Disk Failure Friday Night
https://peakoil.com/forums/server-disk-failure-friday-night-t72855.html
Page 1 of 1
Posted: Sat 24 Sep 2016, 10:53:21Re: Server Disk Failure Friday Night
Posted: Sat 24 Sep 2016, 12:10:21Thanx, wondered f we were under DOS attack again.
Re: Server Disk Failure Friday Night
Posted: Sat 24 Sep 2016, 22:01:32I suppose this is a stupid question, but I'll ask.
Aren't RAID disks super common now? Wouldn't using them eliminate data loss / downtime from a single disk failure?
I worked in systems support in databases for 19 years, so I know all about what happens when a mainframe DASD device (commonly a pair sharing read/write heads) is renedered "dead" when the wrong part fails. (I'd spend 12 to 48 hours doing emergency recovery stuff without sleep).
Virtualizing that DASD with controllers that made it still look like the same devices to the operating system, but having banks of cheap RAID disks hold the actual data completely changed my work life. If a device failed, a little red light came on, the broken disk in the RAID drawer was replaced, the controller recovered it, and I was never even notified.
Just wondered if the issue was cost, or system compatibility with RAID devices, or something else. With these things continuing to crop up, it must be getting mighty painful for the admins. In my experience, an outfit's DASD may grow old together and rapidly become unreliable (relatively speaking -- i.e. lots of failures compared to what's "normal".
FWIW.
As always, I appreciate the efforts of those who make the data recoveries happen, even if at 9 years into retirement, I'm no longer on the leading edge of what goes on in modern databases.
Aren't RAID disks super common now? Wouldn't using them eliminate data loss / downtime from a single disk failure?
I worked in systems support in databases for 19 years, so I know all about what happens when a mainframe DASD device (commonly a pair sharing read/write heads) is renedered "dead" when the wrong part fails. (I'd spend 12 to 48 hours doing emergency recovery stuff without sleep).
Virtualizing that DASD with controllers that made it still look like the same devices to the operating system, but having banks of cheap RAID disks hold the actual data completely changed my work life. If a device failed, a little red light came on, the broken disk in the RAID drawer was replaced, the controller recovered it, and I was never even notified.
Just wondered if the issue was cost, or system compatibility with RAID devices, or something else. With these things continuing to crop up, it must be getting mighty painful for the admins. In my experience, an outfit's DASD may grow old together and rapidly become unreliable (relatively speaking -- i.e. lots of failures compared to what's "normal".
FWIW.
As always, I appreciate the efforts of those who make the data recoveries happen, even if at 9 years into retirement, I'm no longer on the leading edge of what goes on in modern databases.