Re: tables and dbspaces
In article <66pid7$ah4@cssun.mathcs.emory.edu>, Nigel Gall
<nigelg@ppsl.com> writes
>
>At 03:45 PM 12/11/97 EDT, you wrote:
>>
>>I wonder if there is any problem with creating one dbspace per table?
>>We have about 130 tables, and this grouping can help us to monitor the
>>I/O requests per table.
>>I read the guides and I didn't see any warning about applying such a
>>solution.
>>
>1 dbspace per disk implies >1 chunk per disk. This could mean
that >1 cleaner tries to page clean to the same disk. This would be
the disk heads start thrashing back and forth hence decrease
performance.
I would go for 1 chunk = 1 dbspace = 1 disk. Remember to keep the
rootdbs small (I normally use 40Mb). Also 1 temp dbspace per disk
and list them in DBSPACETEMP in your ONCONFIG.
>>Thanks,
>>Ronald Velghe.
>>
>Hi!
>
>Off the top of my head, I think there's very little wrong with doing this,
>if space allows. In fact, it is suggested by the Performance Tuning Guide
>written by Elisabeth Suto as one of the ways DBA's can examine I/O at the
>table level (using tbstat -D in OnLine 5.x).
>
Under Online 7.x you can use sysmaster:systabprof? Something like
that to examine table I/O. Check $INFORMIXDIR/etc/sysmaster.sql and
look for seqscans.
Or onstat -g opn.
>Go brave, I say!
>
>Best regards,
>Nigel
>+-------------------------------------------------------------+
>|Name : Edmund Nigel Gall Tel: (868) 636 3153 |
>|Title : Information Systems Specialist Fax: (868) 679 3770 |
>|Company: Process Plant Services Limited |
>|Address: Atlantic Avenue, Point Lisas Industrial Estate |
>| Point Lisas, Couva, Trinidad & Tobago, W.I. |
>+----- mailto:nigelg@ppsl.com ------ http://www.ppsl.com -----+
>
--
David Williams