Re: Long Fetch times
Re: Long Fetch times
- Subject: Re: Long Fetch times
- From: Chuck Hill <email@hidden>
- Date: Sat, 13 Mar 2010 20:44:04 -0800
On Mar 13, 2010, at 7:54 PM, Jeff Schmitz wrote:
Thanks Dennis,
comments below...
On Mar 13, 2010, at 8:59 PM, Gaastra Dennis - WO Lists wrote:
Some things coming to my mind:
1) Are you using a lot of indices and/or compound indices?
Sometimes when you have large tables, loading those indices the
first time, takes a while. So there is a fine balance between too
many and not enough indices; we have noticed with FB. As such,
after every server restart, we "warm up" the database to get it
going.
I don't have a lot of indices, just the default and one or two others.
That sounds like "not near enough".
The thing is, once it's gone, it's gone. I can restart, restore
from live backup (haven't tried flat files), reboot, doesn't matter,
after a certain, sudden point any fetch takes on the order of
minutes, even to return no data.
It is not how much data it returns, it is how much data it has to sift
though first.
2) If your DB is too fragmented, consider writing it to a flat-
file, and restoring it; as shown in the FB docs.
Would flat-file maybe work better than from live backup?
Yes. But if you problem is lack of indexes / too much data, that
won't help. Also try running
optimize database;
in the SQL pane of FrontBaseManager.
Chuck
3) How is your underlying storage medium doing? Enough free disk
space? Consider deploying on SSDs.
Should be plenty. Honestly, it doesn't take that much data in the
database to get this to happen. A live backup gives a file on the
order of 100 megabytes.
With Kind Regards,
Dennis Gaastra,
Chief Technology Officer,
WEBAPPZ® Systems, Inc.
On 2010-03-13, at 4:44 PM, Jeff Schmitz wrote:
While running some stress tests I seem to be able to get my
database (Frontbase) in a state where fetch times take an
inordinate amount of time (e.g. fetches that return no rows take a
minute), and once in that state, even a reboot of the machine
won't fix the problem. Is there anyway to recover such a
database? I'll be perusing the Frontbase for any ideas, but from
experience, is such behavior symptomatic of any particular
problem? I've been running several years and haven't until now
seen such behavior.
Thanks,
Jeff
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden
--
Chuck Hill Senior Consultant / VP Development
Practical WebObjects - for developers who want to increase their
overall knowledge of WebObjects or who are trying to solve specific
problems.
http://www.global-village.net/products/practical_webobjects
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden