Fix I/O errors on a chunk

If an I/O error occurs during the database server operation, the status of the chunk on which the error occurred changes to down.

If a chunk is down, the onstat -d display shows the chunk status as PD- for a primary chunk and MD- for a mirror chunk. For an example of onstat -d output, see the HCL OneDB™ Administrator's Reference.

In addition, the message log lists a message with the location of the error and a suggested solution. The listed solution is a possible fix, but does not always correct the problem.

If the down chunk is mirrored, the database server continues to operate using the mirror chunk. Use operating-system utilities to determine what is wrong with the down chunk and correct the problem. You must then direct the database server to restore mirror chunk data.

For information about recovering a mirror chunk, see Recover a mirror chunk.

If the down chunk is not mirrored and contains logical-log files, the physical log, or the root dbspace, the database server immediately initiates a stop action. Otherwise, the database server can continue to operate but cannot write to or read from the down chunk or any other chunks in the dbspace of that chunk. You must take steps to determine why the I/O error occurred, correct the problem, and restore the dbspace from a backup.

If you take the database server to offline mode when a chunk is marked as down (D), you can restart the database server, provided that the chunk marked as down does not contain critical data (logical-log files, the physical log, or the root dbspace).