@derobert wow. I'll keep that in mind the next time I encounter a failed selftest, and then I'll try an md resync
@derobert another thing. Will RAID edition hard disks behave differently in this scenario? I assume an RE disk would not keep a sector pending, so that md can handle the situation without a complete check?
@derobert another thing. Will RAID edition hard disks behave differently in this scenario? I assume an RE disk would not keep a sector pending, so that md can handle the situation without a complete check?
A RAID edition HDD should actually generate more pending sectors...
The normal error recovery sequence goes: read sector, notice checksum doesn't match, ecc can't recover, attempt re-read, repeat (several times...)
This takes a while
The RAID disks, at least if you turn on the limited error recovery time, will do a lot fewer rereads
The eventual end of that error sequence is either "finally managed to read sector, attempt to write back correct data, re-read, if not correct reallocate" or "give error to OS, mark sector pending"
And, ideally, upon receiving a read error, the RAID will fairly soon send back a write with the correct data (mdadm does this, for quite a while)
So md would solve this practically on its own, but only if it's actually a read from md that causes the sector to go pending. When it's the selftest, we're back at having to force an md check
We resell servers from a big German computer manufacturer; they are all sold with 24 hours on-site support. When we see that a disk is failing, we send a mail to the manufacturer, and they send one of their local partners to fix it.
@MartinvonWittich I haven't seen a CPU fail yet, either.
Intel and AMD seem able to handle QC. At least for CPUs.
Bad disks, bad optical drives, bad mainboards, bad RAM, bad power supplies, bad video cards, bad cables, bad disk backplanes... but never a bad CPU. Pretty amazing, actually.
Hm. I've never actually run an overclocked system. I can't imagine overclocking a P4 was easy. I mean, those suckers ran hot, even when underclocked....
I'd love to help but my knowledge of NTP can be summed up by 'some service that lets you set the time from internet servers" so can't really contribute much.
@terdon No problem. You can help out by giving some things you've always wondered about NTP. Or that have caused you to curse when the magic didn't work.
I guess what most people (including me) expect from NTP is that it "just works". So I'm especially curious if there's a way to ensure that ntpd will keep the time in sync regardless of big discrepancies between the hardware clock and the current time. I'm looking forward to the parts "why jumps are bad", and then "when you do it anyway" :)
Actually, will manually changing the time (using normal 'date', or NTP or whatever) cause a cron job to run? As long as you don't set it to the actual minute the cronjob is scheduled for, I would not expect it to affect the cron daemon.
@terdon if you have a job schedules for 03:07, and the clock is currently 03:10, it's probably already run. If you set the clock back to 02:57 (because the clock was wrong), that job is going to run again
> So you're missing something else. My crystal ball has tucked away for the night, so if you want help finding the problem in that file, you're going to have to post it. – Gilles 3 mins ago
I'm trying to cut + copy + paste in tmux using 'evilvrt' under debian testing and using xclip, but the situation is compounded by me having an alias set up for xclip, which I've now commented out, but when I've restarted .bash_aliases its failing saying -bash: .bashrc: line 385: syntax error: une...
@Braiam - uh yeah the first sentence: "I am trying to install centos onto another one of my machines, and I am new to Ubuntu. When I downloaded centos the operating system came as two .iso images. So far I have been unable to figure out how to mount the two images onto a single 32gb flash drive to do a full install on the"
Questions describing a problem that can't be reproduced and seemingly went away on its own (or went away when a typo was fixed) are off-topic as they are unlikely to help future readers.
Sorry, that's what I meant, its what 'too localized' changed to in the Great SE Close Vote System revamp.
From the post you link to:
5(A). "Too Localized" is no longer necessary since the specific off-topic reasons now address its main use case. (new)
Too Localized was, by far, the most misused close reason in our surveys, with both Community Managers and Moderators deeming over 50% of rando...
Questions describing a problem that can't be reproduced and seemingly went away on its own (or went away when a typo was fixed) are off-topic as they are unlikely to help future readers.
From the post you link to:
5(A). "Too Localized" is no longer necessary since the specific off-topic reasons now address its main use case. (new)
Too Localized was, by far, the most misused close reason in our surveys, with both Community Managers and Moderators deeming over 50% of rando...
