0

I am having Internet problem with Ubuntu Server.I have installed Ubuntu server in my college server and internet connection type is DHCP. I am getting lights at the backside of the CPU where we insert LAN cable and if i put the same cable to other PC then its getting the Internet with out any problem.

here some error really bothering me 

end_request:I/O error,dev sda ,sector 17548653

I am getting this error in the logs.as I know this sounds to me some drive issue but this error I am getting every time while restarting the network with service command.

help me to solve this issue.

Raja G
  • 102,391
  • 106
  • 255
  • 328
  • 2
    Possibly, some file that is needed for the interface is located in the problematic sector. If you get a message like that, you should shut down the server (or unmount the drive that throws the error if that is possible, which it probably isn't) and check the disk before doing anything else. – January Jul 08 '13 at 10:34
  • Try sudo dhclient and see what happens. – Anders Jul 08 '13 at 19:06
  • Did you have (more) problems with the disk since this happened? (I know this is old, and I understand if you don't have much more info about it.) Often this is part of a pattern of drive failure, though based on the somewhat unusual presentation (being triggered specifically by disconnecting/connecting to a network) and the absence of any other messages from the logs, I'm not totally sure this necessarily was the usual disk failure problem (like in how to interpret these errors from syslog) – Eliah Kagan Sep 09 '14 at 21:50

1 Answers1

0

OK, the comment is too long. My first guess is that a file needed to start the interface is located in a damaged disk sector. You can test that hypothesis with debugfs: Identifying file associated with unreadable disk sector (on gra2.com) is a good explanation how to do it. It's not straightforward, mind you, but maybe it would be worthwhile to get through the steps just to make sure the two phenomenons are really connected.

In short, you need two steps:

  1. convert the LBA (logical block address) of the bad block into the block address of the filesystem. LBAs are counted in different units than the filesystem block numbers, and start relative to the beginning of the disk, whereas the filesystem block numbers start at the partition start
  2. once you have the filesystem block number, you can use debugfs to find the inode associated with that particular block. Hopefully you will find that (a) the file or files cannot be accessed and (b) have something to do with setting up the interface.

The above is not the remedy but forensics -- understanding what is going on.

However, I'd suggest to first shut down the system, boot it from the CD and do all subsequent operations from the rescue system. A damaged disk is always a bad sign

guntbert
  • 13,134
January
  • 35,952