20040704: EXT2 errors on bigbird's RAID

>From: Unidata Support <address@hidden>
>Organization: UCAR/Unidata
>Keywords:  200406220437.i5M4beWb016431 LDM RAID storage IDD CRAFT

Hi Gerry,

I have been meaning to send you a note ever since bigbird was restarted
after the A&M power outage, but I kept forgetting.

Two issues:

- I thought that you had left the RAID formatted with a 'jfs' file
  system.  It appears to be EXT2 now

- bigbird is seeing lots and logs of EXT2-fs errors.  Here is a snippit
  from /var/log/messages.1:

Jul  2 15:36:22 bigbird kernel: EXT2-fs error (device md(9,0)): ext2_new_block: 
Free blocks count corrupted for block group 10117
Jul  2 15:36:22 bigbird kernel: EXT2-fs error (device md(9,0)): ext2_new_block: 
Free blocks count corrupted for block group 472
Jul  2 15:36:22 bigbird last message repeated 12 times
Jul  2 15:36:22 bigbird kernel: EXT2-fs error (device md(9,0)): ext2_new_block: 
Free blocks count corrupted for block group 7338
Jul  2 15:36:22 bigbird kernel: EXT2-fs error (device md(9,0)): ext2_new_block: 
Free blocks count corrupted for block group 472
Jul  2 15:36:22 bigbird last message repeated 3 times
Jul  2 15:36:22 bigbird kernel: EXT2-fs error (device md(9,0)): ext2_new_block: 
Free blocks count corrupted for block group 7338
Jul  2 15:36:22 bigbird kernel: EXT2-fs error (device md(9,0)): ext2_new_block: 
Free blocks count corrupted for block group 472
Jul  2 15:36:23 bigbird last message repeated 7 times
Jul  2 15:36:23 bigbird kernel: EXT2-fs error (device md(9,0)): ext2_new_block: 
Free blocks count corrupted for block group 10117
Jul  2 15:36:23 bigbird kernel: EXT2-fs error (device md(9,0)): ext2_new_block: 
Free blocks count corrupted for block group 472
Jul  2 15:36:23 bigbird last message repeated 40 times


So, did you reformat the RAID as EXT2?  If yes, then I think that the
LDM should be stopped, the RAID unmounted, and fsck run on the RAID to
clean up bad blocks.  If the RAID is still supposed to be formatted
with JFS, then something bad has happened to its setup.

>From: Gerry Creager N5JXS <address@hidden>
>Date: Wed, 30 Jun 2004 11:14:26 -0500
>Organization: Texas A&M University -- AATLT
>To: Tom Yoksas <address@hidden>
>Subject: CRAFT action question for generating images using gpnexr2

>Per our recent discussion, I'm trying to sort out the best way to do the 
>action on receipt of the entire volume scan for a radar site.  The 
>example below is based on the Amarillo, TX site.

# CRAFT
CRAFT   
^L2-BZIP2/(....)/([0-9][0-9][0-9][0-9][0-1][0-9][0-3][0-9])([0-2][0-9][0-5][0-9])([0-9][0-9]).*/p(...)(KAMA)
 
        PIPE    decoders/dcnexr2 -s \1 -d /dev/null 
        data/gempak/nexrad/craft/\1/\1_\2_\3
        EXEC    bin/KAMA-L2.sh

>where bin/KAMA-L2.sh is the script to fire off gpnexr2 with the 
>appropriate parameters?  Am I on the right track?

As I said in our phone conversation, you need a separate pattern/action
for each distinct action.  Also, the regular expression you include will
not match the CRAFT headers.  Your pattern is looking for the '/p......'
kind of pattern that is available for the Level III products.

So, the above needs to be changed to:

# CRAFT
CRAFT   
^L2-BZIP2/(KAMA)/([0-9][0-9][0-9][0-9][0-1][0-9][0-3][0-9])([0-2][0-9][0-5][0-9])([0-9][0-9])
        PIPE    decoders/dcnexr2 -s \1 -d /dev/null 
        data/gempak/nexrad/craft/\1/\1_\2_\3

# CRAFT
CRAFT   
^L2-BZIP2/(KAMA)/([0-9][0-9][0-9][0-9][0-1][0-9][0-3][0-9])([0-2][0-9][0-5][0-9])([0-9][0-9])
        EXEC    bin/KAMA-L2.sh
        data/gempak/nexrad/craft/\1/\1_\2_\3

>I'm gonna be on 
>vacation and trying this while we drive (wireless modems have some 
>benefits), but I don't intend to do a lot of phone calling, really...

You shouldn't!

>I've killed all other scripts and the resource problems appear to have 
>calmed somewhat.  Still, the right way to do this will be to start doing 
>scripts, where possible, as pqact actions.

>That won't work, of course, for Level III mosaics, but that's another 
>issue.  Neither, I believe, for our countours for temp, dewpoint, heat 
>index, or pressure.  Those, however, only run once an hour.

>Thanks, Gerry


--
NOTE: All email exchanges with Unidata User Support are recorded in the
Unidata inquiry tracking system and then made publically available
through the web.  If you do not want to have your interactions made
available in this way, you must let us know in each email you send to us.