 |
chained fetch ratio |
29 march 2000 |
in my statistics i am getting chained fetch ratio 0.3099 pctfree too low for a tableis there any way i can find out which table has pctfree too low? | ||
 |
this could be chaining, rather than row migration. if it is row migration, and if your schema is analyzed, then our row_migration.sql script might help. |
|
block headers |
10 april 2000 |
| i'm starting to take some blockdumps, but it is hard to know what i'm looking at. can you help me to understand what all those header fields are used for? they seem to take up quite a lot of space. | ||
|
there are three sets of header information, corresponding to the three layers of the oracle kernel that handle data blocks. i'll only explain the first two for now, because the third is too big a question to answer. firstly, all datafile blocks are written and read by the cache layer (kcb) generally through the database block buffer cache. there is a 20-byte header and 4-byte tail that are read and maintained by the cache layer. the cache layer header is also called the common block header.
blocks that contain user data are further interpreted and maintained by the transaction layer (ktb). these blocks have a transaction layer header immediately following the common block header. the transaction header is comprised of a 48-byte fixed header, including one 24-byte interested transaction slot, and a variable number of additional interested transaction list slots. the remainder of each block, the data area, is used by the data layer (kd) to store user data. the data area contains a header, free space and row data. the internal structure of the data layer header and row data area is dependent on the block type. cluster blocks, table blocks, index branch blocks and index leaf blocks each have different stuctures. as i said, i'll not attempt to explain the structure of this area in detail for now, but i will tell you how big the headers are. here's how it looks in a diagram.
the cache header and tailthis is what the cache header and tail look like in a block dump.buffer tsn: 0 rdba: 0x004000f8 (1/248) scn: 0x0000.0001ab63 seq: 0x01 flg: 0x00 tail: 0xab630601 frmt: 0x02 chkval: 0x0000 type: 0x06=trans data the cache layer header is comprised of the following fields.
the tail is comprised of the low-order two bytes of the scn base followed by the block type and the sequence number. the consistency of the header and tail is checked whenever a block is read. this detects most block corruptions, in particular split blocks from hot backups. the transaction headerthe transaction header consists of 24 bytes of control information followed by an array of one or more transaction slots. this array is called the interested transaction list (itl), and each slot may contain an interested transaction entry (ite). the initial number of itl slots is set by the initrans parameter for the segment, except that index branch blocks only have one slot. cluster and index segments have a minimum initrans value of 2. if there is sufficient free space in the data area of the block, additional transaction slots can be dynamically allocated as required, up to a limit dictated by the database block size or the maxtrans parameter for the segment. this is what the transaction header looks like in a block dump. block header dump: 0x00400035 object id on block? y seg/obj: 0x2 csc: 0x00.21ab6 itc: 2 flg: - typ: 1 - data fsl: 0 fnx: 0x0 ver: 0x01 itl xid uba flag lck scn/fsc 0x01 xid: 0x0000.00f.00000005 uba: 0x0040000b.0000.12 c--- 0 scn 0x0000.000000ac 0x02 xid: 0x0002.003.00000027 uba: 0x00c00004.0019.01 --u- 1 fsc 0x0000.00021b7c the control information in the transaction header contains the following fields.
|
