Data Block Dump in Oracle – part 2 – how to read block dump file
Sometime you look into dump/trace files in USER DUMP DESTINATION, aka udump, generated by Oracle to dig into issues, or you dig into problem by dumping some Oracle blocks, you may see something like
Start dump data blocks tsn: 4 file#: 4 minblk 172 maxblk 173
buffer tsn: 4 rdba: 0x010000ac (4/172)
scn: 0×0000.16b676dc seq: 0×01 flg: 0×04 tail: 0x76dc0601
frmt: 0×02 chkval: 0x78c7 type: 0×06=trans data
Object id on Block? Y
seg/obj: 0x2dca4 csc: 0×00.16b676dc itc: 3 flg: E typ: 1 – DATA
It is essential to understand who is who. Here I will explain a few attributes in the dump file, highlighted in BOLD.
rdba is the Data Block Address, which can be translated into FILE NUMBER and BLOCK ADDRESS using DBMS_UTILITY.DATA_BLOCK_ADDRESS_FILE to get the file number part of a data block address and
DBMS_UTILITY.DATA_BLOCK_ADDRESS_BLOCK to get the block number part of a data block address.
SCN is the block’s SCN, together with flg: etc for transaction control and block cleaning.
seg/obj is the DATA_OBJECT_ID, which pointing to the seg$ location of the block.
type means the type of data block, where “trans data” means table data block or index block – coded as 0×06. There are more than 40 types, like 0×20 “FIRST LEVEL BITMAP BLOCK”. For more, you can visit “block types” by Julian