Table does not exist even if there is one

Question

I have MySQL master-slave replication setup in Amazon EC2. For the sake of resilience, I have set up LVM on RAID 10 array, on both master as well as slave. I have been using ec2-consistent-snapshot for some time now to take the snapshot from the slave instance.

So, I tried to create the disks from those snapshots and attach them to a new MySQL installation. Since the new installation would not recognise the LVM and RAID setup, I reactivated the LVM and RAID in the new instance using the following commands.

1. Scan the disks for RAID array members:

   mdadm --examine --scan <comprising disks>
   

2. Reactivate the RAID array:

   mdadm -A -s
   

3. Reactivate the lvm:

   vgchange -ay
   

Since snapshotting took the backup of the ibdata and iblogfiles, I removed them for the fresh installation. MySQL would start normal.

Help me out of this

Answer 1

Removed ibdata1? You can* remove iblogfiles but not the ibdata1!! They're system tablespaces storing tables, indexes, metadata, undologs....

*=> Removing iblogs is okay once you stop mysql but not while it's running!

Update: Removing iblogs is okay only after shutting down mysql with innodb_fast_shutdown disabled.

You might like to review this answer to understand what is wrong! You might want to do a restore from the snapshot may be.