Geez. As if I needed another problem... harrier.slackbuilds.org, which provides (among other things) mirrors.slackware.com, slackbook.org, SBo mailing lists, rlworkman.net, and DNS for those and quite a few others, is down for the moment.

I rebooted harrier earlier tonight to bring up 3.10.107, and the main OS drive is now showing as an "Unsupported device" by the server, so I have nice remote console that's absolutely useless. This drive was a brand new (as of March 2013) Dell-branded enterprise drive, and there were no indications of problems prior to the reboot, so I don't really know what happened.

I'll try to get up to the datacenter tomorrow, but it might not happen until Thursday or even Friday.

The good news is that all of the useful data on it is in a raid array on the other three drives, and all of those *appear* to be fine.

UPDATE: we're back up, albeit in a non-ideal state. Everything is working, but I don't trust the replacement drive either. More work is planned soon.

This server also hosts my blog at http://alien.slackbook.org/blog/ ... so that one is down as well.

2 members found this post helpful.

Count https://sbopkg.org as well

2 members found this post helpful.

for now sbopkg.org is up again after i moved it temporarily to another server

2 members found this post helpful.

My favorite setup is a 2 drives RAID1 for OS, and to use 3 (or 4) drives on RAID5 for data (i.e. sites)

Not trying to give lessons to gurus, but you know... shit happens.

And a RAID1 for OS have also the advantage of that distributed reading, just like RAID0, then a better responsiveness of the system.

1 members found this post helpful.

We used to have exactly that (back in 2011 when the system was first brought online), but with two drives RAID1 for OS and two drives RAID1 for data. In 2013, the two data drives died, and since the other two were from the same batch, I decided to put the data (the important stuff) on three drives and leave only one for the OS itself. The OS drive was a brand new (well, it had a couple of months of burn-in and had proven (I thought) to be reliable), so it seemed reasonable at the time.

At this point, the only way to go back to a more ideal setup (2+2) would be a complete reinstall and lots of time at the datacenter. I don't have that kind of time, nor do I want to burden the DC guys since they're providing the colo for free, so we'll take our chances for another couple of years once I get the new drive (it won't arrive until Monday) installed. I also ordered a vflash module for the idrac, so we'll be in a little better shape going forward just in case...

3 members found this post helpful.

On a side note, yesterday the french provider Online launched a special offer for root servers. I got myself a bare metal Xeon Quad Core with 16 GB RAM, 2x2 TB hardware RAID 1 drives, full root access and unlimited bandwidth for less than 20 euros a month and no extra installation fees. I've been working with Online since 2013 (six servers so far), and their quality of service is outstanding.

Cheers,

Niki

2 members found this post helpful.

I had an incredibly poor experience with online.net and would not recommend them.
They gave me a broken machine that had disc errors, apparently caused by a known problem.
In a nut shell - they repeatedly told me that I should pay more money for a higher spec machine (that I didn't need) in order to deal with a problem that was
a) known and they didn't bother to check the service before provisioning
b) that was exclusively theirs.

I hope you have a better experience!

2 members found this post helpful.