Hi Ken,
Sorry I may not have got you. So isn't there a workaround for the
issue with **existing* *records? I think I need not change DB since
I've already compiled MySQL with 'sphinxse'. It is the sphinx
configuration I'm concerned with.
Thanks,
On Monday 04 February 2013 10:33 PM, [email protected] wrote:
On Mon, Feb 04, 2013 at 10:28:23PM +0530, Subin wrote:
Hello all,
Thanks to Arkadiusz Miskiewicz for sharing his knowledge here:
http://lists.fsck.com/pipermail/rt-users/2011-August/072464.html
My question is based exactly on this config which I find is good.
"As per my understanding" the 'delta' index makes newly created
rows/records available by using 'SphinxCounters' table. That is
the delta index which will be run every 15 minutes will check if
any row which has 'id' greater than the 'id' stored previously
in 'max_doc_id' column of 'SphinxCounters' table. If greater
then it will read the source and then add the new record to the
'delta' index.
If an edit/update happens to an *"existing"* ticket/document/row in
the source database, how will this change get updated into indexes
quickly? So according to this config, we will need to wait until
the next time the 'main' is reindexed again, right? If true how
can we have a workaround for this problem?
Please excuse and let me know if I made any wrong assumptions
here.
Thanks,
Hi Subin,
That is how the Sphinx full-text indexing works. You would need to
change to a different DB backend with integrated support for fulltext
indexing. PostgreSQL and Oracle will both work. Eventually, when MySQL
supports full-text indexing with InnoDB natively, RT can be made to
use it as well.
Regards,
Ken
