After added some Chunks, My Informix can't perform level 0 backup!!

[porra]

My server is K420,HP-UX 10.2
Recently,a HP engineer help us to added 5 new Chunks into my system. However, now my Informix Database can't perform level 0 backup.

The problem is:
$Ontape -s
Please enter the level of archive to be performed(0,1,or 2)0
DBspace datadbs_a is disabled.
Program Over!

In the same time,I can perform logic log backup.

This is my Chunk informiation of Informix:
$ onstat -d

Informix Dynamic Server Version 7.30.UC3 -- On-Line -- Up 07:48:42 -- 266952 K
bytes

Dbspaces
address number flags fchunk nchunks flags owner name
cbf43158 1 1 1 1 N informix rootdbs_a
cbf43f78 2 1 2 9 N informix datadbs_a
cbf44038 3 1 6 1 N informix phydbs_a
cbf440f8 4 1 7 1 N informix logdbs_a
cbf441b8 5 2001 8 1 N T informix tempdbs_a
5 active, 2047 maximum

Chunks
address chk/dbs offset size free bpages flags pathname
cbf43218 1 1 0 150000 147811 PO- /informix/informix/ro
otdbs_a
cbf434f8 2 2 0 1000000 21 PO- /informix/informix/da
tadbs_a1
cbf435d8 3 2 0 1000000 5 PO- /informix/informix/da
tadbs_a2
cbf436b8 4 2 0 1000000 5 PO- /informix/informix/da
tadbs_a3
cbf43798 5 2 0 1000000 120937 PO- /informix/informix/da
tadbs_a4
cbf43878 6 3 0 125000 4947 PO- /informix/informix/ph
ydbs_a
cbf43958 7 4 0 400000 24947 PO- /informix/informix/lo
gdbs_a
cbf43a38 8 5 0 350000 349947 PO- /informix/informix/te
mpdbs_a
cbf43b18 9 2 0 1000000 0 PD- /informix/informix/da
tadbs_a6
cbf43bf8 10 2 0 1000000 0 PD- /informix/informix/da
tadbs_a7
cbf43cd8 11 2 0 1000000 0 PD- /informix/informix/da
tadbs_a8
cbf43db8 12 2 0 1000000 0 PD- /informix/informix/da
tadbs_a9
cbf43e98 13 2 0 1000000 0 PD- /informix/informix/da
tadbs_a10
13 active, 2047 maximum

I need your help!!

[Rainer von Bongartz]

Please also post the output of

onstat -d


Regards
Rainer

[Nick Wickens]

I guess that the 5 new chunks are those showing as PD in the onstat -d output. Check that the volumes for the chunks are still available and any file permissions/ownerships match the other chunks (You should not have been able to add the chunks if they were wrong but worth checking.) Failing that get the engineer back.

[Steve Lewis · Accepted solution]

1. Ensure your volume group is activated and your new disks are still visible. Did you bounce the server recently?

2. Check the online log file to see why it took the chunks off-line in the first place. Then work backwards from this, checking that everything is in place, as it should be - the chunk files(or links), the permissions, the logical volumes, the volume group, the disks. Then fix those things first.

3. Is this a MC/Servieguard cluster? If so, then
if your package switched nodes and you hadn't re-vgimported your volume group into this node then informix will show raw (lvol) chunks as being down because it wouldn't see the logical volumes. The solution in that case, is to vgimport the volume group to the current node, set the permissions of the chunks to 660, chown informix:informix, start up then engine and change the status of each chunk using
onspaces -s datadbs_a -p /informix/informix/datadbs_aX -o 0 -O

(thats -o [zero] -[capital] O on the end.)

If you are brave and have a backup taken before you added the chunks you could try forcing the chunks online anyway. Only do this if you are entirely sure that everything is OK at the HP-UX level and you have a backup.