Status: Verified
SYMPTOM(s):
proutil dbname -C dbanalys gets bkioRead errors every time the command is issued.
<function>:Insufficient disk space during <system call>, fd <file descriptor>, len <bytes>, offset <bytes>, file <file-name>. (9450)
(9450) bkioRead:Insufficient disk space during Read, fd 5244, len 8192, offset 229376, file dbname\data_extent_number
SYSTEM ERROR: read wrong dbkey at offset <offset> in file <file> found <dbkey>, expected <dbkey>, retrying. area <number> (9445)
(9445) SYSTEM ERROR: read wrong dbkey at offset 469762048 in file dbname\data_extent_number
The block dumped in the database log file (dbname.lg) is a bad block. Header information is wrong.
The block dumped using DBRPR after a fresh reboot has the correct header information.
The 9450 errors change when the command is reran.
DBRPR utility reporting significant amount of "Can't find block messages"
Can't find block messages start at block 131072
These errors also change when the command is ran more than once.
proutil dbname -C dbrpr reporting Bad Frag XXXXXXX for rec XXXXXXX Err:3 for a significant amount of records.
No errors being reported while database is in a multi-user mode.
Able to use 4GL program to read all records within area that DBRPR is reporting errors against.
proutil dbname -C dbtool to validate records and blocks are not reporting any errors.
proutil dbname -C idxcheck reported 35 errors where it cannot find a key.
SYSTEM ERROR: Index<index-num> (<owner-name>.<table-name>, <index-name>): couldn't find key <key> recid <DBKEY>. (8783)
proutil dbname -C idxfix deleted and then fixed 70 errors.
Index <index number> (<owner name>.<table name>, <index name>): Deleted key <key value> recid <record identifier>. (8828)
Index <index name> (<owner name>.<table name>, <index name>): Added key <index value> recid <record identifier>. (8827)
Users are reading and writing to database without receiving any errors.
The dbanalys option to proutil was reporting errors until prostrct builddb was used to recreate the dbname.db file.
Hardware has been scanned with no errors.
No file system errors being reported in system logs.
Areas with Can't find block messages are Type II storage areas.
Tools such as dbanalys, idxfix and DBRPR are showing corruption that appears to be changing every time the commands are ran. In other words, different blocks or different areas reporting errors every time the commands are issued.
FACT(s) (Environment):
Database has an 8 KB database block size.
Windows
OpenEdge 10.1B03 Service Pack
CAUSE:
Problem was caused by a cache issue on the machine (file system or controller). Once machine was rebooted, the errors no longer occurred.
FIX:
Identify and address the cache issue with the machine.
A workaround for the issue can sometimes be to reboot the machine. This will result in cache being cleared out for the time being so that commands can be used to validate the database. If errors persist, there may be a need to disable the cache until the underlying issue can be addressed. Contact hardware or Operating System vendor for additional assistance.