Can the damaged non-system hard disk bring down server?
The answer is absolutely yes！ The following article showthe scenario I met.
Environment:
Dell 2950, 2 local SATA disks.
One is 250G, with ESX 4.1 system, the other is 500G, onlydata.
One day, I found some of my vms are grey out under one server in VI inventory. First thing in my mind is: Shit! Hard disk is broken. Becausethey’re local disks, not ISCSI, if this situation occur, 90% it is.
Check the Summary of the ESX server from vSphere Client, twoDatastores show, 250G is health, 500G is grey out.
Just have some high priority work need to do, leave it, planto handle it at noon.
When I come back from my work, the situation changed, not “everythingunder control” now.
Both the vms on that server are grey out, both 2 datastoresgrey out, the server goes down.
What’s happened? I rushed into our lab.
Found that: The power of Dell 2950 is still on, blackdisplay of monitor, keyboard also no response.
Then I try to rescue our server:
A. Reboot it, but not work, still hang.
B. Try to enter BIOS during reboot, no response.
C. Unplugthe power, wait for some time, reconnect power and reboot, try BIOSagain. This time I am “lucky”, the monitor display some vertical stripsinstead of black.
D. Disconnect both of the 2 disks, the server enterBIOS successfully.
E. Connect only 250G, ESX system startup well.
F. Connect 500G together, server hang during startup.
Until now, I can tell that this damaged 500G hard disk bring down the server.
PS: Later I try to connect this damaged 500G disk to windows system, not successful.
The label of this disk is marked “Guarantee to 11/5/28”.
