We seem to be having some issues with incomplete rebuilds of of Category table. We currently have 1300 Categorys (these have been modified to allow subcategories in the description) and the past two days, the build has stopped at or about entry 1260. We don't get a complete build report (ie 268 seconds). What are/is the issue here?

If you are doing it by http, you might be exceeding the system time out, or settings for running processes. Try it again by telnet.

If you are on a shared server, your processes might be killed by a daemon that prevents any one process from using too much CPU or run time.

If you are on a dedicated server, run it by telnet, and see what the error messages are. You need to hit a ctrl-D to get past the CGI.pm name/value pair prompt.

http://www.postcards.com
FAQ: http://www.postcards.com/FAQ/LinkSQL/

Hello!

I have had severe problems and you may look in to my threads untill last year.

Finally I gave up as the climate became very hot and people started saying that its only me who had this stupid problems.

From what you say, I know this. You may not waste your time and wait untill Alex brings some Genius solutions. Do not try telnet as it would be also a waste of time.

The build routines are a complete and most troublesome routines and are a disaster, even though one must end up saying that they are really GOOD programming codes. If the base of build routines are not good enough, the entire product becomes less worth and unsable as it happened to me. Alex is to blame for this attitude that since one year I am not able to start my website in a good way.

Hi,

I have posted a build staggered modification in the resources. I hope it will help all who have problems with the building-time / cpu usage.

regards alexander

it can be downloaded

http://www.silberling.com/lsql/staggerd.txt

Alexander,

I quess you maked a lot of people happy !:)....

Keep up the good work!

Regards Startpoint..

alexander,

Just installed your mod - brilliant - thanks for all your hard work :))

All the best
Shaun

alexander,

Could you do a similar thing for nph-index.cgi?

I've used your nph-build.cgimod to add over 5000+ categories, but now my 'Re-Index' is timing out occasionally - could you modify it so it works in stages and indexes x records at a time just like your build-mod?

All the best
Shaun

Ok,

here it is

The howto manual http://www.silberling.com/...to_nph-index_stg.txt

the new cgi script http://www.silberling.com/lsql/nph-index_stg.txt

please notice that it is an ugly hack, increasing the indexing-time 5x (not calculating the pausing time), and using far to many SQL calls.. but it works

regards, alexander

Alexander,

Thanks for the mod, but it seems to cause problems; it takes a lot longer than a normal re-index and it doesn't appear to index correctly, or does something to the indexes that Links doesn't like.

Here's the error I got when trying to search after doing a re-index staggered:

This is the sub sub fetchrow_arrayref but I'm not that good at perl myself so I can't offer any help on where the problem might be :(

If I use the standard re-index afterwards, it returns to normal and behaves as I'd expect.

All the best
Shaun

Hi Shaun,

sorry for the malfunction. I am just hacking Search.pm
in order to get the f... (?#§$&) AND & OR to work
and can't use it at the moment.

please try to upload the file again. I made some changes.
Please inform me about any errors.

regards, alexander

Hi alex,

Sorry, got the same errors :(

Again it took a long while to do the actual indexing (avg. 400-500 records per 15 secs.) and then procudes the search.pm error after a long pause when clicking 'search'.

All the best
Shaun

Hi,

Sorry that the mod is not working on your server, it works on mine!?

Maybe I got some time later for it.

Im still in Search.pm.....

regards, alexander

I know rajani's message is from more than a month ago, but I must ask: Do you stick around here for any other reason than to bad mouth Links SQL? You've said repeatedly that Links SQL did not work for you. Guess what, millions of people have had products not work for them. And the majority of those products do not receive nearly the support that Gossamer Threads provides. It's a simple concept -- if it doesn't work for you, move on.

Having read many of those messages from last year, I don't recall anything about the climate being heated other than your demands to have things tailored to suit your needs. At the time, I was rather shocked at how patient Alex was with the situation.

The advice given was very solid. Telnet is often the solution to builds that time out. Why would you suggest someone not take a few seconds to see if it works for them? Are you really that down on the product that you would rather people throw up their hands in despair and wait a year for someone to come to their rescue?

Alex is in no way responsible for your web site. That responsibility is upon you and you alone. If I were him, I would be very offended by your words and demand an apology. What Alex is responsible for is producing a product that works well for as many people as possible, and I think the vast majority of Links SQL users would agree that he has done a great job with that thus far.

You have other choices for how to go about your web site, be it competing products or a totally different direction. Do not blame others for the time you have needlessly wasted.

Dan

 
This MOD is really really really helping me.
Alexander... thanks!!

Regards