ssosw
asked on
Server room Airconditioning failed
Hi,
The A/C in our server room failed and we were unaware of it for about two hours. Actually I don't worry about it because there is a stand alone unit as well as the central cooling so if one is off then the other is cooling the room . This morning the Central system was shut down for maintenance( of which i was informed very late) and at the same time( or earlier) the stand alone unit decided to quit. Dont know what caused it. I notice a couple of hours later that the server room(3.5 by 1.6 mts with two racks/ 6 servers /10KVA ups/ switches no ventilation except for the A/C's, Glass sliding doors) is very hot .The room was cooler after opening the doors but there seems to be some damage done on one of the servers. One HDD on this machine ( RAid 5 with 5 SCSI HDD's) is showing a red indication with a cross.The other 4 HDD's are ok with a green light. I checked server and see no problems .Everything is working fine. I checked the Disk MAnagemnet and it says the Logical drive is healthy. I need to know where I can find if something is wrong. Also what does the red cross on the HDD mean( obviously it is still working )Is this caused by the temprature rise or could it be something that was already there and I missed it.
Also the other servers ... what possibilitise are ther that there is damage caused due to the temprature. Is it treu that tempratures upto 60C are bearable by the equipments.Is there anything that I can setup that will monitor the temperature on servers / Server room and notify me.
I also need to know if the servers will shut down if the temperatures rise above limits( before getting damaged) . Is there a way to do this .Would my servers have shutdown if it had continued for another two hours in the same condition. The thing is I do not know what damage the servers took and need some idea of what maight have hapenned.
The A/C in our server room failed and we were unaware of it for about two hours. Actually I don't worry about it because there is a stand alone unit as well as the central cooling so if one is off then the other is cooling the room . This morning the Central system was shut down for maintenance( of which i was informed very late) and at the same time( or earlier) the stand alone unit decided to quit. Dont know what caused it. I notice a couple of hours later that the server room(3.5 by 1.6 mts with two racks/ 6 servers /10KVA ups/ switches no ventilation except for the A/C's, Glass sliding doors) is very hot .The room was cooler after opening the doors but there seems to be some damage done on one of the servers. One HDD on this machine ( RAid 5 with 5 SCSI HDD's) is showing a red indication with a cross.The other 4 HDD's are ok with a green light. I checked server and see no problems .Everything is working fine. I checked the Disk MAnagemnet and it says the Logical drive is healthy. I need to know where I can find if something is wrong. Also what does the red cross on the HDD mean( obviously it is still working )Is this caused by the temprature rise or could it be something that was already there and I missed it.
Also the other servers ... what possibilitise are ther that there is damage caused due to the temprature. Is it treu that tempratures upto 60C are bearable by the equipments.Is there anything that I can setup that will monitor the temperature on servers / Server room and notify me.
I also need to know if the servers will shut down if the temperatures rise above limits( before getting damaged) . Is there a way to do this .Would my servers have shutdown if it had continued for another two hours in the same condition. The thing is I do not know what damage the servers took and need some idea of what maight have hapenned.
On linux servers I use http://www.lm-sensors.org/ for that purpose. I'm sure there equivalent for Windows and others.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
>One HDD on this machine ( RAid 5 with 5 SCSI HDD's) is showing a red indication with a cross.The other 4 HDD's are ok with a green light. Everything is working fine. I checked the Disk MAnagemnet and it says the Logical drive is healthy.
The raid is probably degraded (one failed disk) but from a logical view still working fine. Windows disk management can't see the failed drive, but the raid diagnostics tool from your vendor will allow you to confirm that the disk has failed.
You need to replace the failed drive asap to avoid server failure should another disk fail.
Yes, replace the failed disk with the made and model and same or bigger size. Allow 4 to 6 hours for RAID 5 to rebuild itself. Make sure you replace the drive while server is UP and RUNNING so RAID 5 can be rebuilt.
K
K
The drive may have its own temperature sensor in which case it might have turned itself off rather than failed. In that case you may get away with simply reseating it and it will rebuild into the array. Manufacturer's diags may be able to interrogate it still and see if that is the case.
ASKER
Hi,
The server is an HP Proliant ML570.
It shows on the Compaq Integrated Management Log Viewer and entry that says
Drive Array Device Failure (Slot 1 Bus 1 Bay 4)
The thing is the date is 17th Nov . The temprature rise was last thursday.
The server is an HP Proliant ML570.
It shows on the Compaq Integrated Management Log Viewer and entry that says
Drive Array Device Failure (Slot 1 Bus 1 Bay 4)
The thing is the date is 17th Nov . The temprature rise was last thursday.
The drive has indeed failed and should be replaced ASAP.
BTW: this server has an integrated ILO management processor. Is this isn't already configured, you really should. If you combine this with HP Insight Manager software, you can setup an early warning system that sends e-mail when certain temperature levels are exceeded.
Either CMOS clock is incorrect or the disk failed weeks ago and you didn't notice. That's the problem with redundancy, if you do not monitor the system you may have failures and not notice them.