Hi,
we have a server that runs on redhat linux. It was working fine until last year and all of sudden we are not able to ssh on to that server. we are getting the ping response but not getting the login prompt when tried to connect via ssh.
Port 22 is opened and i have tested it out using telnet command.
So i have requested the guys at the data center to reboot the machine forcefully and when he tries to do it nothing come up on the screen it seems.
When its powered on nothing appears on the screen seems like it is completely dead and the keyboard doesn't work it seems. There is no boot process when powered off/on.
Any advise how to proceed ahead so that we can boot the server and resolve the issue?
Thanks,
P
It's a bit late, but whenever you can't get to a system via the network, a better next step is to get to it via the system console. You could have asked the
guys in the data center to do this.
"Nothing comes up on the screen" is too vague.
What kind of machine is this? W
What is attached to the system console of the machine? Are you using KVM? Does this server have a service processor?
Do you get a bios screen at least? Can you go into the setup of the bios at the system console?
Hi,
Here is more info. this is work station on which guava software is running to monitor all the oracle databases running on other servers.
It is a IBM machine and only IBM logo is comming up it seems on the screen and keyboard is not responding to any key strokes as well as mentioned by datacentre guys.
It is running on RHEL 4 and it doesn't have any console access which is bad.
So for the time being i have requested the data centre guys to reboot the machine and see if they can find any errors if so asked them to send a screen shot as we dont have console access to it.
Any ideas how to proceed ahead?
Thanks,
P
---------- Post updated at 01:56 PM ---------- Previous update was at 11:01 AM ----------
Hi,
Update after requesting the data center guys to restart the machine they came back saying that boot sequence stops when it starts reading system from HDD.
Currently they are trying to clone HDD it on another and see what it gives.
Any suggestions how to proceed next?
Thanks,
P
It looks like the boot disk is faulty.
So trying to clone it is a good idea.
If that is successful, replace the disk by the clone disk.
Furthermore, (if not successful,) an expensive SCSI disk has a possibility to "repair" a sector (remap to a spare sector). The contents of a "repair"ed sector is unknown, that can mean a little data loss/corruption. (That's why everyone says: re-install the OS, then restore from last data backup.)
Hi,
Unfortunately the cloning the hard disk has failed and we dont have the backups of this machine as this is a desktop which is a mistake to have on it without backups.
At the moment i thinking of two ways
-
use rescue mode by booting from installation media and see if we can manually mount the hard disk, if so recover the scripts and application data software that is needed.
-
Connect this hard disk as a secondary hard disk to any other linux server and see we can mount it or not?
Not sure how to proceed on repairing the sector?
Thanks,
P
Yes you can do 1. and 2.: trying to mount, or running fsck on the device first then mount.
2.: use a non-production Linux system for this! If there is HW error (retrying to read sectors) the system will partly or fully freeze for a few seconds or minutes (dependent on the strategy in the kernel drivers).
I have no experience how to handle the repair of sectors on x86 systems!
I did this several times on Sparc systems (format -> analyse -> read test), so the disk became readable again. Such a "repair"ed disk will get more bad sectors, so it is advisable to immediately clone the disk.
Need to see how far these can work.
The guys at the data center doesnt have any linux knowledge and that is really hard to explain them how it should be done.
So waiting for one of the technical guy who comes back to work mostly this weekend which will be more helpful in recovering the data.
Anyway thanks for the suggestions and will keep it posted how it goes.
If any one has any more ideas please contribute.
Thanks,
P