PDA

View Full Version : Red Server Status


Glorindale
07-27-2013, 10:25 PM
Anyone know when the server might be back up?

Nads
07-27-2013, 10:29 PM
Probably when it's ready.

Gustoo
07-27-2013, 10:36 PM
Says late sunday probably. Rogean tired.

Glorindale
07-27-2013, 10:36 PM
ah bollocks. Oh well.

Tomato King6
07-27-2013, 10:36 PM
drive array is busted he will need a new array not just the one or two disks and then it has to be copied over

days , most likely 3-5

Genedin
07-27-2013, 10:38 PM
I really think its amazing that two separate entire drive arrays on two separate boxes broke at the same time. That's some really ineradicable odds.

runlvlzero
07-27-2013, 10:43 PM
drive array is busted he will need a new array not just the one or two disks and then it has to be copied over

days , most likely 3-5

I've done raid setups and clones b4, hours at most. Probably 40 mins on good modern hardware. And not ancient 7200 rpm drives

I really think its amazing that two separate entire drive arrays on two separate boxes broke at the same time. That's some really ineradicable odds.

I wouldn't be surprised if red went down to fix blue.

If boxes were hosted on rented hardware, would have been fixed in minutes. A complete failure would have been avoided. It takes 3 drives to go bad in a raid 5, before things start getting ugly. The people hosting the hardware would have thrown drives in it.

These are Basement Dweller Servers though. Lets not deceive ourselves.

Glorindale
07-27-2013, 10:46 PM
I really think its amazing that two separate entire drive arrays on two separate boxes broke at the same time. That's some really ineradicable odds.

That's not what happened. There were two drivers on the same array. If you lose two drivers on lets say a raid 5 array you volume is toast. It is unusual to have two drive failures on an array in such a short period of time which makes me wonder if the first hard driver failure occurred some time ago but wasn't detected (you can operate with one failed drive but take a performance hit) until the second one failed which brought the volume offline. This is why you always replace a failed hard drive ASAP.

Glorindale
07-27-2013, 10:51 PM
I've done raid setups and clones b4, hours at most. Probably 40 mins on good modern hardware. And not ancient 7200 rpm drives



I wouldn't be surprised if red went down to fix blue.

If boxes were hosted on rented hardware, would have been fixed in minutes. A complete failure would have been avoided. It takes 3 drives to go bad in a raid 5, before things start getting ugly. The people hosting the hardware would have thrown drives in it.

These are Basement Dweller Servers though. Lets not deceive ourselves.

It sounds like they have some servers in a locker at an ISP.

Two drivers in a raid 5 will bring it down, not three. Which is why you always replace a failed hard drive asap.

runlvlzero
07-27-2013, 10:51 PM
That's not what happened. There were two drivers on the same array. If you lose two drivers on lets say a raid 5 array you volume is toast. It is unusual to have two drive failures on an array in such a short period of time which makes me wonder if the first hard driver failure occurred some time ago but wasn't detected (you can operate with one failed drive but take a performance hit) until the second one failed which brought the volume offline. This is why you always replace a failed hard drive ASAP.

Two separate boxes. Raid 5 can degrade to a lower raid with less performance backup 1 drive to a 2ncd.

The odds are pretty astronomical on two separate servers that both would degrade to a critical level where a failure would mean data loss and they'd need to both be brought down and fixed for integrity.

Unless they were already failing like you stated. But both already failing is really bad from a sys admin perspective. Shit needs fixed when it goes bad.

If things were setup right there'd be a drive in a box ready to get slid into a tray and hot plugged right on.

Days is a dumbass estimate unless u have to build an entirely new array and wait on hardware. Even rebuilding an array from one working drive doesn't take very long. Its pretty much text file edit and hardware swapping. If the OS and raidsetup is already done. You just set the working drive to be the master and plug it into the new array and bamb. Nothing special needs be done.

Bazia
07-27-2013, 10:52 PM
take it to tells nerds

Genedin
07-27-2013, 10:57 PM
This was my point. They have stated that these are two separate boxes and the problem with both of them were the arrays failing........
It is pretty amazing to me not only these odds, bu that they wouldn't have a drive on standby to hot-swap in.

You cant run an array like this for years and just expect that a drive will never go out.
Are we just assuming this is a RAID-5 setup?

Glorindale
07-27-2013, 10:58 PM
I don't think there were two separate server failures. I think it was one array on one server, the blue server. I think they may have cannibalized the red servers hard drives in order to get the blue server back up. I could be wrong though.

Raid 5 will run with one failed hard drive but not two.

Maybe the long delay is due to shipping of the new hard drives.

Glorindale
07-27-2013, 11:01 PM
This was my point. They have stated that these are two separate boxes and the problem with both of them were the arrays failing........
It is pretty amazing to me not only these odds, bu that they wouldn't have a drive on standby to hot-swap in.

You cant run an array like this for years and just expect that a drive will never go out.
Are we just assuming this is a RAID-5 setup?

I was just guessing. Raid 5 would be the cheapest if you wanted a lot of spindles. However, I don't know what they got going on.

I am not suprised they didn't have a hot spare. Remember, this is a not for profit operation not a cooperation.

runlvlzero
07-27-2013, 11:01 PM
Correct were assuming its raid5, worst case scenario u can make raid five raid 1 with 2 working drives still. Meaning 2 drive failures with 4 drives.

Drives are cheap. Any extra left over in the budget of the project aught to have gone to 1-2 spares. Like 90$ at most.

Unless project is running at a deficit.

Glorindale
07-27-2013, 11:02 PM
take it to tells nerds

so says the someone playing a 13 year old MMORPG.:rolleyes:

Loli Pops
07-27-2013, 11:06 PM
Probably after he wipe completes. Takes hours.

Pudge
07-27-2013, 11:06 PM
ya........... my theory all along. take from red to prop up blue.

wish they kept red alive to let blues try us out for a spin. they know reds ain't blue'in, cept to cause trouble (solb trains)

Colgate
07-27-2013, 11:10 PM
rogaine does not give a fuck about red or any of you

the only reason he would bring it back up at all so his public image of running "stable" servers isnt tarnished

just be glad he's at least an ego maniac

inb4 ban

Nirgun
07-27-2013, 11:54 PM
velious firiona vie is best firiona vie

Glorindale
07-28-2013, 11:12 PM
So is tonight the night or no?