Very Computer

# metadb
# metaclear -a
metaclear: biostat: d6: stale databases

#

Why am I getting this stala database error if there are no databases
left (AFAIK)?

--
Mauricio                                raub-kudria-com      
(if you need to email me, use this address =)

# metadb -a -f -c2 c2t0d0s0 c2t0d0s1
# metadb
        flags           first blk       block count
     a        u         16              1034            /dev/dsk/c2t0d0s0
     a        u         1050            1034            /dev/dsk/c2t0d0s0
     a        u         16              1034            /dev/dsk/c2t0d0s1
     a        u         1050            1034            /dev/dsk/c2t0d0s1
# metaclear -a
metaclear: biostat: d6: stale databases

#

--
Mauricio                                raub-kudria-com      
(if you need to email me, use this address =)

1) Save the configuration

2) Mount the devices as c0t0d0sX and reboot

3) Create a new set of state databases

4) Mirror the devices from scratch

There may be an easier method depending on how things
are currently oriented, but this will allow you to
start fresh. You might want to save the output of
"metastat -p" in case this happens again.

Quote:> 2) Mount the devices as c0t0d0sX and reboot

Quote:> 3) Create a new set of state databases

Quote:> 4) Mirror the devices from scratch

Quote:> There may be an easier method depending on how things
> are currently oriented, but this will allow you to
> start fresh. You might want to save the output of
> "metastat -p" in case this happens again.

Quote:

>> I would presume becuase the configuration is stored in the
>> meta state database. If you try to remove items w/o the
>> configuration, how would metaclear know what to kill? Someone
>> please correct me if I am wrong.
>    Idf that is the case, what could I do?  I could try creating new
> database replicas, but that does not seem to take me very far:
> # metadb -a -f -c2 c2t0d0s0 c2t0d0s1
> # metadb
>         flags           first blk       block count
>      a        u         16              1034            /dev/dsk/c2t0d0s0
>      a        u         1050            1034            /dev/dsk/c2t0d0s0
>      a        u         16              1034            /dev/dsk/c2t0d0s1
>      a        u         1050            1034            /dev/dsk/c2t0d0s1
> # metaclear -a
> # metaclear: biostat: d6: stale databases

--

Senior Technical Consultant         TAOS            http://www.taos.com/
Got some Dr Pepper?                           San Francisco, CA bay area
         < This line left intentionally blank to confuse you. >

Since I am running Solaris8 in this E450, I have tried running metatool and
then deleting the two metadevices (I was going to replace both for a single
larger setup).  All seems to be well when I delete them, but if I quit
metatool and then restart it, the two metadevices will be there waiting for
me.  Why would that happen?  Could it be that even metatool is acting up
(which would not surprise me given the history of how this machine was
managed) and is not updating whatever it has to (/etc/system comes to
mind).

Quote:>> created the new databases?  It looks like something's wedged in the
>> running DB.  Sounds like a bug.

>            I have been thinking to be the case myself.  And, here are my
> concerns:  let's say the metadevices are corrupted beyond salvation,
> causing the stale databases.  Is there a file I can go and kill any
> reference to the metadevices so when I reboot the machine has no idea they
> were ever around?  

After that, you have no replicas and the system doesn't know of any.
You therefore have no SDS configuration at all.  You'll need to create
new replicas and initialize any metadevices.

Quote:> Since I am running Solaris8 in this E450, I have tried running metatool and
> then deleting the two metadevices (I was going to replace both for a single
> larger setup).  All seems to be well when I delete them, but if I quit
> metatool and then restart it, the two metadevices will be there waiting for
> me.  Why would that happen?  Could it be that even metatool is acting up
> (which would not surprise me given the history of how this machine was
> managed) and is not updating whatever it has to (/etc/system comes to
> mind).

Check /etc/system to see if it has references to metadb locations.

--

Senior Technical Consultant         TAOS            http://www.taos.com/
Got some Dr Pepper?                           San Francisco, CA bay area
         < This line left intentionally blank to confuse you. >

Have you tried the "-d" option for metadb(1M)?

     -d    Deletes all replicas that are located on the specified
           slice.   The /etc/system file is automatically updated
           with the new information and the /etc/lvm/mddb.cf file
           is updated.

A friend of mine sent the following notes he had:

    To clear a drive from the meta database:

        metaclear -r d17
        metadb -d mddb17

    The following will clear one drive from the state replicas

        metadb -d c3t9d0s0

--
David Magda <dmagda at ee.ryerson.ca>, http://www.magda.ca/
Because the innovator has for enemies all those who have done well under
the old conditions, and lukewarm defenders in those who may do well
under the new. -- Niccolo Machiavelli, _The Prince_, Chapter VI