Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations strongm on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

heat failure

Status
Not open for further replies.

lunatic23

Technical User
Mar 26, 2003
63
GB
hey guys,

the bottom line is that our unix server crashed yesterday, i'm a total newbie when it comes to unix but it was squealing about "unrecoverable errors reading SCSI drive..." & "non system disk or disk error"

now our server room isn't well suited for servers & there are occasional crashes due to the heat build up im the room, and on the unix server when this happens it shows "Non system disk or disk error" messages, and after ten minutes cooling period its up and running again.

what im asking is, could the heat cause permanent problems to the hard drive, ie: permanent damage to boot records? physical damage to plates?

right now the drive is off to a recovery team who wil hopfully get our data back.

thanks for any input on this situation
luni

someone once told me to shut up...
 
Of course it can!
Mechanically, due to unnecessary expansion caused by heat mya stress some mechanical parts and change their specs. That will cause damage.

On the other hand, lubricants in the spindle ball bearing evaporate at high temperatures, so you will again have problems.

Due to the expansion of the surface and of the read/write head arm, data will be written in a marginal part of the track, and can be difficult to read back when cold.

Electronics will also suffer from high temperature and the timing may change and/or become marginal.

Over heat is over risk.

But judging by the fact that "experts trying to recover the disk", I understand that your company loves risk. Did you ever hear something called "backup", or "redundancy", or "raid" (though I'm not sure if Unix supports raid - it should)

One actual trouble is better than a thousand advices.
 
If you can work there without taking out of room breaks every 10 minutes the server should be able to work there without problems.
But if you have airflow problems in the Unix box, or CPU fan problems, you can make the problems worse with a hot room.
More likely the HD was coming to the end of its life.

Sounds like time for a "unix administration" course.

Ed Fair
Any advice I give is my best judgement based on my interpretation of the facts you supply. Help increase my knowledge by providing some feedback, good or bad, on any advice I have given.
 
Actually Servers should be in an environmentally controlled room,and yes rooms of 80 degrees can cause problems, Fans are great,but if all they are doing is moving hot air around it does no good, the object is to exhaust hot air and bring into the server box cooler air. And some unix boxes and servers can have up to 13 fans and if 1 doesn't work,then neither will the server. Also some manuf. will bill for repairs caused by environmental problems,I know. I have done it.

And yes, Raid will work with unix,I have set it up many times on IBM rs/6000 systems running AIX (IBM UNIX). But Raid is not a catch all, you still should use at the very least a tape backup with off site storage. What would happen to your data if you had a fire,flood,tornado,or any other calamity. If you have the money for an outside source to try to recover your data,you have the money to backup that data properly.
 
Best advice.... GET A DROP with some MAJOR AIR FLOW coming in there. Its like this, There is absolutely nothing that heat can't damage. Every Server room I have even been in has been kept atleast 74 or below. I used to work for IBM/Comm Data. They kept their Server room at 68. It was an ice box. But things ran forever.

Even if you don't destroy Hard Drives and burn up CPU's, you will eventually burn the board. Those chips aren't made for that kind of abuse. Same principle applies with the cold. If you turn the computer off at night and it gets down to say 40 or even less. The board gets very cold and then in the morning you flip on all those currents of electricity, it is just like throwing hot water on a iced over windshield. Eventually it will crack and so will your chips. If you ever let it get so cold that it frost and then hit the juice......buzzzzzzzz.....smoke.....lights....camera....action.....FRIED.

ME
 
thanx guys,

i really don't take anything to do with the unix server, all i do is change the backup tapes each day [yep there is a backup] & im not sure what the motive was for sending it off to the repair people...im guessing cuz the drive was screwed and there was nowhere to restore the backup to at that point.

the bottom line is that the disk is 'unrecoverable'. its a big fat paperweight now. but my boss managed to get a hold of an old HP unix server, and he's trying a restore from tape.

with regards to the heat issue:
- the room is workable, tho it gets a bit stuffy in the summer
- there is an air-conditioning unit in place, which blows cold air onto the servers,
- the average temp of the room during summer is 27-29 degrees C, if it goes any higher than 30 the servers start to shutdown,
- and to help matters the central heating boiler is in the same room, top marks for the planning there...

thanks for all the replies, its certainly opened my eyes to a few things


rclarke250, a small insight to my wonderful employers:
they can afford to blow unnecessary money on "repair experts" but wont buy a half-decent tape drive to backup the Exchange server & they wont do anything about the 50k free space left on the file server boot drive, and the fact that there's no virus protection cuz "that doesn't matter just now": these are minor issues, but he has to prove to mr director that he is doin something about the unix issue...

i'm just the kid who gets shouted at every so often


thanks again
luni


someone once told me to shut up...
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top