we managed to bring online 3 chunks out of 4 using onspaces
command.
One chunk (datdbs7) has this error when we
tried to bring it online:
"ISAM Error": Primary and mirror chunks are bad"
-----------------
The history is that we have added 2 hard disks and
created 6 new chunks in October, 2003. We do not
have any problem until lately when we discovered
that 4 chunks went down. So now we have brought up
3 chunks and 1 chunk is still down!
Don't know if this chunk has data, the 3 chunks that
were successfully online are empty.

-----------------
Greatly appreciate any inputs, suggestions to solve this
problem.



Informix Dynamic Server Version 7.30.UC9 -- On-Line -- Up 18:01:20 -- 193680 Kbytes

Dbspaces
address number flags fchunk nchunks flags owner name
c5f0b158 1 1 1 1 N informix rootdbs
c5f0c918 2 1 2 1 N informix plogdbs
c5f0c9d8 3 1 3 2 N informix llogdbs
c5f0ca98 4 2001 4 1 N T informix tmpdbs
c5f0cb58 5 1 5 10 N informix datdbs
c5f0cc18 6 1 6 10 N informix idxdbs
6 active, 2047 maximum

Chunks
address chk/dbs offset size free bpages flags pathname
c5f0b218 1 1 0 100000 91413 PO- /informix/rootdbs/rootdbs
c5f0b4f8 2 2 0 50000 49947 PO- /informix/plogdbs/plogdbs
c5f0b5d8 3 3 0 100000 947 PO- /informix/llogdbs/llogdbs
c5f0b6b8 4 4 0 250000 249947 PO- /informix/tmpdbs/tmpdbs
c5f0b798 5 5 0 1000000 7 PO- /informix/datdbs/datdbs
c5f0b878 6 6 0 500000 1 PO- /informix/idxdbs/idxdbs
c5f0b958 7 5 0 1000000 3 PO- /informix/datdbs/datdbs1
c5f0ba38 8 6 0 1000000 70 PO- /informix/idxdbs/idxdbs1
c5f0bb18 9 5 0 1000000 5 PO- /informix/datdbs/datdbs2
c5f0bbf8 10 6 0 500000 1 PO- /informix/idxdbs/idxdbs2
c5f0bcd8 11 3 0 100000 48997 PO- /informix/llogdbs/llogdbs1
c5f0bdb8 12 6 0 500000 21241 PO- /informix/idxdbs/idxdbs3
c5f0be98 13 5 0 1000000 53 PO- /informix/datdbs/datdbs3
c5f0bf78 14 6 0 1000000 202269 PO- /informix/idxdbs/idxdbs4
c5f0c058 15 5 0 1000000 312285 PO- /informix/datdbs/datdbs4
c5f0c138 16 6 0 1000000 552774 PO- /informix/idxdbs/idxdbs5
c5f0c218 17 5 0 1000000 264853 PO- /informix/datdbs/datdbs5
c5f0c2f8 18 6 0 1000000 100945 PO- /informix/idxdbs/idxdbs6
c5f0c3d8 19 5 0 1000000 43949 PO- /informix/datdbs/datdbs6
c5f0c4b8 20 5 0 1000000 0 PD- /informix/datdbs/datdbs7
c5f0c598 21 5 0 1000000 0 PD- /informix/datdbs/datdbs8
c5f0c678 22 6 0 1000000 686433 PO- /informix/idxdbs/idxdbs7
c5f0c758 23 6 0 1000000 0 PD- /informix/idxdbs/idxdbs8
c5f0c838 24 6 0 1000000 0 PD- /informix/idxdbs/idxdbs9
c61244b0 25 5 0 1000000 999997 PO- /informix/datdbs/datdbs9
25 active, 2047 maximum

First problem :
Informix Dynamic Server Version 7.30.UC9 is no longer supported by IBM (End of Service =31-Mar-03).

Are the chunks symbolic links to the disk space? If yes, do the links exist and underlying lvol exist?

Other options include doing a restore from before the chunks were marked down, then applying logical logs forward. If you don't care what is in the chunk, I suppose you could try to drop the chunk and readd it.

You could search http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&group=comp.databases.informix

Andy

hi Andy,
Thank you very much for your suggestion.

We are unable to drop the chunk either.
Is there a way where we can see what's inside this chunk.
We were able to bring online all the chunks except for
datdbs7. it says "ISAM error.................."

I don't know of a way to see what is in a certain chunk, you can only see what is in a dbspace.

I made a mistake earlier by suggesting to drop the chunk - since it has data in it, Informix will not allow you to drop it.

The only 2 possible solutions I see:
1. Call informix support: see if they will mark the chunk online
2. Replace the chunk either by replacing the disk or by linking another disk to the same path or link. Then restore the entire dbspace from backup.


Andy

Hi,

We had a similar experience recently.

A chunk was flagged as bad. We tried a restore first but there was a log-tape problem. As we are on 7.23ucl we are not officially supported but still maintained the contract as we are going to migrate very soon.
I contacted Informix support (they were brilliant!) who managed to bring the chunk online for us. A recovery server was used to restore as recent a database as possible. We then used unload/dbload to move missing data back to the repaired tables.

Paul