CFINDEX stops responding

[Rich_Willmore]

We are trying to index with TYPE=PATH a directory which contains roughly 20k
files. These files are a collection of HTML, DOC, and PDF. I have tried
multiple times to index this directory using CFINDEX. Though according to the
task manager the index has completed in roughly 20-30 minutes the CFINDEX CFM
template does not respond. Furthermore my post CFINDEX processing tags do not
fire off either. I have ommitted all file extensions except HTM/HTML and still
get the same result. However, in the end the index is healthy and I can bounce
criteria off of it with CFSEARCH. It does seem intact. Any suggestions? I
would like to schedule this CFM template to run nightly but fear that we will
run into problems since the post CFINDEX processing tags deletes certain
black-listed keys from this particular site -- very imporant. -- Rich

[Rich_Willmore]

Any help??

-- Rich

[Tom Jordahl]

What version CF?

Does the index work if you break the work in to smaller chunks? Is there a
particular file or directory that causes the template to get stuck?

--
Tom Jordahl
Macromedia Server Development

[robstacey]

I have the same problems with indexing if anything else is happening on the
server at the same time. So, what I do is run an initial, manual index on the
really large collections from the administrator last at night (when I know
there will be no activity) and then have a sceduled task to run every night
that is designed to update all the documents. I am sucessful updating about
50k documents a night with this method.

[Rich_Willmore]

We are CF 6.1 w/ latest patches etc... The CFINDEX does not seem to stop on a
particular file or file type. I do have better luck when indexing smaller
chunks of the data with CFINDEX.

Indexing via CF admin does seem to work better and updating the index on a
scheduled interval would pick up all the newly added documents. The problem
(to the best of my knowledge) is deleting the KEY(s) from the index once a file
has been removed. I do not believe UPDATE will do that -- may be wrong. So
scheduling a nightly REFRESH was going to be our approach.

Many many thanks for the help thus far.

-- Rich

[robstacey]

Refresh will purge then update the collection, thereby deleting the keys and
readding them. Update will just add the new keys. So, if you have a lot of
changes in what you are indexing then yes, refresh is better, but will take
considerably longer to accomplish than update would. Just as a point of
reference, to run an update on all my collections takes 4-5 hours a night.

[trjlove]

Try
<cfset L_displayvolume = Trim(Replace('#L_displayvolume#', Chr(34), "XXXX", "All"))>

[CoachChip]

<cfset L_displayvolume = Trim(Replace(L_displayvolume, Chr(34), "XXXX", "All"))>

Tried the above suggestion. Didn't work. Any other ideas? I could work
around it but some of my input files have the quotes and some don't.

Thanks...

[shawnwindler]

CoachChip,

I had to do this before (a few months back), so this is from memory, which is
rather shoddy nowadays.

I believe I used the escape character rather than Chr(34).

Something like Replace(myString, '\"', "XXXX", "All") I think should work.
Single quote, backslash, double quote, single quote.

- Lemme know if that works.

- Shawn