Re: page header structure in 9.40

Posted in 2004

Uday Kale — Thu, 15 Apr 2004 09:16:41 -0500 — source: Usenet: comp.databases.informix

--0__=09BBE4E4DFDD97F08f9e8a93df938690918c09BBE4E4DFDD97F0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: quoted-printable New page header is still 24 bytes for >=3D 9.40. We took out 4 bytes of= pg_stamp, and replaced it with a 2-byte checksum over the page header fields and the page-ending timestamp, and the rest 2-bytes were used fo= r the chunk#, old pg_addr is now page offset. Regards, Uday. owner-informix-list@iiug.org wrote on 04/15/2004 05:24:26 AM: > Hi, > > what is the exact structure of the page header in 9.40 ? > > Versions < 9.40: > we have 24 byte for the header including 4 byte page address > (3 nibbles(12 bit) chunk#, 5 nibbles(20 bit) page_offset ) > pg_addr 4 Byte > pg_stamp 4 > pg_nslots 2 > pg_flags 2 > pg_frptr 2 > pg_frcnt 2 > pg_next 4 > pg_prev 4 > > Version 9.40: > I created a chunk larger 2 Gbyte > (more than 2**20 =3D 1.048.576 pages possible ) > oncheck -pP shows 24 Byte page header ( slot 1 starts at 24 )> where and how are chunk# and pageoffset stored or what is > missed in comparison to the old (9.30) format ??? > > TIA > > Ruediger Papke= --0__=09BBE4E4DFDD97F08f9e8a93df938690918c09BBE4E4DFDD97F0 Content-type: text/html; charset=US-ASCII Content-Disposition: inline Content-transfer-encoding: quoted-printable <html><body> New page header is still 24 bytes for >=3D 9.40. We took out 4 by= tes of pg_stamp, and replaced it with a 2-byte checksum over the page h= eader fields and the page-ending timestamp, and the rest 2-bytes were u= sed for the chunk#, old pg_addr is now page offset. Regards, Uday. <tt>owner-informix-list@iiug.org wrote on 04/15/2004 05:24:26 AM: > Hi, > > what is the exact structure of the page header in 9.40 ? > > Versions < 9.40: > we have 24 byte for the header including 4 byte page address > (3 nibbles(12 bit) chunk#, 5 nibbles(20 bit) page_offset ) > pg_addr    4 Byte > pg_stamp   4 > pg_nslots   2 > pg_flags   2 > pg_frptr   2 > pg_frcnt   2 > pg_next      4 > pg_prev      4 > > Version 9.40: > I created a chunk larger 2 Gbyte > (more than 2**20 =3D 1.048.576 pages possible ) > oncheck -pP shows 24 Byte page header ( slot 1 starts at 24 ) > where and how are chunk# and pageoffset stored or what is > missed in comparison to the old (9.30) format ??? > > TIA > > Ruediger Papke </tt></body></html>= --0__=09BBE4E4DFDD97F08f9e8a93df938690918c09BBE4E4DFDD97F0-- sending to informix-list

↪ replying to Uday Kale

Ruediger Papke — 16 Apr 2004 00:13:32 -0700 — source: Usenet: comp.databases.informix

Uday Kale <udayk@us.ibm.com> wrote in message news:<c5m6re$mjk$1@terabinaries.xmission.com>... > --0__=09BBE4E4DFDD97F08f9e8a93df938690918c09BBE4E4DFDD97F0 > Content-type: text/plain; charset=US-ASCII > Content-transfer-encoding: quoted-printable > > > > > > > New page header is still 24 bytes for >=3D 9.40. We took out 4 bytes of= > > pg_stamp, and replaced it with a 2-byte checksum over the page header > fields and the page-ending timestamp, and the rest 2-bytes were used fo= > r > the chunk#, old pg_addr is now page offset. > > Regards, > Uday. Now we have only 1 timestamp. Right? Thanks for your answer Ruediger > > owner-informix-list@iiug.org wrote on 04/15/2004 05:24:26 AM: > > > Hi, > > > > what is the exact structure of the page header in 9.40 ? > > > > Versions < 9.40: > > we have 24 byte for the header including 4 byte page address > > (3 nibbles(12 bit) chunk#, 5 nibbles(20 bit) page_offset ) > > pg_addr 4 Byte > > pg_stamp 4 > > pg_nslots 2 > > pg_flags 2 > > pg_frptr 2 > > pg_frcnt 2 > > pg_next 4 > > pg_prev 4 > > > > Version 9.40: > > I created a chunk larger 2 Gbyte > > (more than 2**20 =3D 1.048.576 pages possible ) > > oncheck -pP shows 24 Byte page header ( slot 1 starts at 24 )> > where and how are chunk# and pageoffset stored or what is > > missed in comparison to the old (9.30) format ??? > > > > TIA > > > > Ruediger Papke= > > --0__=09BBE4E4DFDD97F08f9e8a93df938690918c09BBE4E4DFDD97F0 > Content-type: text/html; charset=US-ASCII > Content-Disposition: inline > Content-transfer-encoding: quoted-printable > > <html><body> > New page header is still 24 bytes for >=3D 9.40. We took out 4 by= > tes of pg_stamp, and replaced it with a 2-byte checksum over the page h= > eader fields and the page-ending timestamp, and the rest 2-bytes were u= > sed for the chunk#, old pg_addr is now page offset. > > Regards, > Uday. > > <tt>owner-informix-list@iiug.org wrote on 04/15/2004 05:24:26 AM: > > > Hi, > > > > what is the exact structure of the page header in 9.40 ? > > > > Versions < 9.40: > > we have 24 byte for the header including 4 byte page address > > (3 nibbles(12 bit) chunk#, 5 nibbles(20 bit) page_offset ) > > pg_addr    4 Byte > > pg_stamp   4 > > pg_nslots   2 > > pg_flags   2 > > pg_frptr   2 > > pg_frcnt   2 > > pg_next      4 > > pg_prev      4 > > > > Version 9.40: > > I created a chunk larger 2 Gbyte > > (more than 2**20 =3D 1.048.576 pages possible ) > > oncheck -pP shows 24 Byte page header ( slot 1 starts at 24 ) > > where and how are chunk# and pageoffset stored or what is > > missed in comparison to the old (9.30) format ??? > > > > TIA > > > > Ruediger Papke > </tt></body></html>= > > --0__=09BBE4E4DFDD97F08f9e8a93df938690918c09BBE4E4DFDD97F0-- > > sending to informix-list

Glossary

chunk