ASR1004 kernel error message

[enseibert]

Hello

I am getting a strange error message on a ASR1004 and I am not able to find someone else who has had the same issue or any reference to a bug that might be causing the problem. The IOS version is :

Cisco IOS Software, IOS-XE Software (X86_64_LINUX_IOSD-ADVENTERPRISEK9-M), Version 15.1(3)S4, RELEASE SOFTWARE (fc1)

The error message I am receiving is:

Feb 19 2015 10:50:26.629 GMT: %IOSXE-2-PLATFORM: R0/0: kernel: EXT2-fs error (device sda1): ext2_readdir: bad page in #2686977

These are coming through every minute or so. With reference to the kernel would rebooting it solve the problem or would an image upgrade be necessary as it was compiled 1st Aug 2012 and could be a potential bug.

Thank you.

[Iulian Vaideanu]

The internal hard-drive might have a problem...  try this:

asr1#fsck harddisk:
/dev/sda1: clean, 65/9764864 files, 313165/19521301 blocks (check after next mount)
fsck completed with no errors

fsck of harddisk: complete

[enseibert]

Thanks for the response. Is that something that can be tried while the router is operational? Will it interfere with the router and stop it forwarding?

[Iulian Vaideanu]

No, it shouldn't - I just ran it on a production unit.

It's very likely that the harddisk: only stores tracelogs and core files, so even if bad sectors are found and repaired by fsck, nothing critical is affected (the IOS image is probably stored in bootflash:)...

[enseibert]

Thanks for that, it does look like an issue with the hard-drive :(

[Iulian Vaideanu]

Did you run fsck?  What did it output?

[enseibert]

I did and it produced an error message and has taken the hard-drive offline. Unfortunately the message itself has spooled out and cannot be viewed again. The option to fsck the hard-drive is no longer an option.

[Iulian Vaideanu]

Nothing relevant retrievable by "show logging" after the fsck?

If the physical device is still there (for example, if "show harddisk: all" only complains about the filesystem and doesn't say something like "no such device") you could also try "format harddisk:" to attempt to rebuild an usable filesystem there...

[enseibert]

No, nothing was registered in the log when it went off-line. The option to 'show harddisk' has disappeared but is still present in the secondary router with the working disk.

It just appears to have shut it down completely and removed any option for me examine or do any further investigation with it.

[Iulian Vaideanu]

You could also try to "request platform hardware filesystem harddisk: online" to see what happens...  found this while reading about how to replace ASR1K RP HDDs:

http://www.cisco.com/c/en/us/td/docs/routers/asr1000/install/guide/asr1routers/asr1higV8/asr1FRUs.html#pgfId-1060675

[enseibert]

Thanks for that, however it does look like it needs replacing as I get this in the log when running that command:

Feb 20 2015 11:17:32.837 GMT: %IOSXE-5-PLATFORM: R0/0: kernel: sd 0:0:0:0: [sda] READ CAPACITY failed
Feb 20 2015 11:17:32.837 GMT: %IOSXE-5-PLATFORM: R0/0: kernel: sd 0:0:0:0: [sda] Sense not available.
Feb 20 2015 11:17:32.837 GMT: %IOSXE-3-PLATFORM: R0/0: kernel: sd 0:0:0:0: [sda] Asking for cache data failed
Feb 20 2015 11:17:32.855 GMT: %IOSXE-0-PLATFORM: R0/0: udev: harddisk: has been removed
Feb 20 2015 11:17:35.838 GMT: %IOSXE-5-PLATFORM: R0/0: kernel: sd 0:0:0:0: [sda] READ CAPACITY failed
Feb 20 2015 11:17:35.838 GMT: %IOSXE-5-PLATFORM: R0/0: kernel: sd 0:0:0:0: [sda] Sense not available.
Feb 20 2015 11:17:35.838 GMT: %IOSXE-3-PLATFORM: R0/0: kernel: sd 0:0:0:0: [sda] Asking for cache data failed

[Iulian Vaideanu]

Oh well...  at least is it an RP2, so that you can replace the drive without taking the whole router down? :)

[Bradley Marshall]

This worked for my problem! Thanks!

router# dir

Directory of bootflash:/

No files in directory

1339412480 bytes total (704286720 bytes free)

router#

router#fsck bootflash:

/dev/bootflash1 contains a file system with errors, check forced.

Pass 1: Checking inodes, blocks, and sizes

Pass 2: Checking directory structure

Pass 3: Checking directory connectivity

Pass 4: Checking reference counts

Pass 5: Checking group summary information

Free blocks count wrong for group #5 (19734, counted=19476).

Fix? yes

Free blocks count wrong (188556, counted=188298).

Fix? yes

Free inodes count wrong for group #5 (15075, counted=15074).

Fix? yes

Free inodes count wrong (166096, counted=166095).

Fix? yes

/dev/bootflash1: ***** FILE SYSTEM WAS MODIFIED *****

/dev/bootflash1: 49/166144 files (6.1% non-contiguous), 143933/332231 blocks

fsck completed with no errors

fsck of bootflash: complete

router#

[biao li]

最后怎么解决的？