I/O error Err#5 EIO

UNIX Solaris
1

Hi gurus,

My Solaris version is
Solaris 10 1/06 s10s_u1wos_19a SPARC

Im getting I/O error while trying to ls -l a directory in Solaris SPARC box

the truss output from server is attached below

truss -fall -vall -wall -rall ls -la

12278/1:        acl("./dat", GETACLCNT, 0, 0x00000000)          = 4
12278/1:        lstat64("./sh", 0xFFBFFCA8)     (sleeping...)
12278/1:        lstat64("./sh", 0xFFBFFCA8)                     Err#5 EIO
./sh12278/1:    write(2, " . / s h", 4)                         = 4
: 12278/1:      write(2, " :  ", 2)                             = 2
I/O error12278/1:       write(2, " I / O   e r r o r", 9)               = 9

12278/1:        write(2, "\n", 1)                               = 1
12278/1:        lstat64("./dump", 0xFFBFFCA8)                   = 0
12278/1:            d=0x05280004 i=126525 m=0040755 l=2  u=1035  g=103   sz=512

Can you share your suggestions for correcting this error.
Thanks in advance.

2

Here are some suggestions:

  • Have a look to the status of the disk storing the file system containing this directory. (df -k; iostat -En)
  • Be ready to replace it.
  • Make sure you have up to date backups available.
  • Try fsck
  • Update/Upgrade your Solaris version which looks to be nine years old.
  • Switch to ZFS.
1 Like
3

Thanks jlliagre,

iostat is indeed showing errors.

c4t4d0          Soft Errors: 0 Hard Errors: 23526 Transport Errors: 4606
Vendor: SEAGATE  Product: ST373207LSUN72G  Revision: 0354 Serial No: 1030MHGL
Size: 73.40GB <73400057856 bytes>
Media Error: 19634 Device Not Ready: 0 No Device: 3892 Recoverable: 0
Illegal Request: 0 Predictive Failure Analysis: 1597

I have no experience in using fsck.
Im planning to run fsck like this.

1) umount the disk
2) run fsck as 
     fsck /dev/dsk/c4t4d0s0

Hopefully fsck will give me some interactive fixes.
Do you see any issues in running fsck like this and are these hard errors fixable by fsck?

Thanks

4

I think jlliagre was being kind. Disks have what is called an MTBF - which is sort of an expected operating lifetime in operating hours, or uptime. Consider that your disk may be 'fixed' in the very short term by fsck.

But the disk will continue on its path to complete degradation. Your other disks may also have issues.

We had an old SPARC box - and several replacement disks on hand for this very reason. If you do not have support getting older disk hardware becomes really problematic. We have actually bought parts on Ebay. Our benighted management finally got the message - we now have support: full support.