Indexing does not work anymore

[phil@infomed.org]

Hi again,

It looks like everything is working properly now, thanks for your assistance in this

I've asked the hosting co to explicitly allow the execution of scripts in the /cgi-bin/ directory and that seems to have done the trick. The htaccess redirect is now switched back on and the results are as per expectation. Indexing seems to be working properly as well.

Just to be clear, we could run the search under SSL even before, just not when the redirect was on. Somehow the redirect in htaccess and the (missing?) permissions on the directory broke the output.

Is it fair to say that the error had nothing to do with the Zoom version?

Thanks
ph

[Ray]

In which case, we go back to the points for this particular problem from before:

Problem 2:
Your search results render as gibberish when redirect to SSL is enabled. [Shown in screenshot attached to your most recent post.]

Answer 2:

• We can see in the screenshot that, first of all, you have a PHP page (infomed_search.php). I presume this is executing a CGI, capturing the output and displaying it on the page (presumably like the wrappers described in our support page here. Let us know which method is being used, and if otherwise).
• There is quite possibly complications with redirections to SSL and whether PHP is set up to execute a CGI when running under SSL.
• The "gibberish" actually looks like a binary dump of the 'search.cgi' file. This means it is failing to execute the CGI and your PHP code is causing it to simply dump the contents of the CGI file.
• Interestingly, going to www.infomed.ch/cgi-bin/search.cgi causes the CGI file to download (with either http or https). This means it isn't setup to be executed properly. However, I can't confirm if this is the actual CGI file you are calling from infomed_search.php, or a stray copy from a previous setup attempt.

[...]
The second problem is very unlikely to be fixed by upgrading to V7. It is almost definitely a web server configuration issue.

Please verify the first point and if or how your infomed_search.php is executing the CGI.
Please verify the location of your CGI file, and try to launch this directly via your browser via both http:// and https://

If the above indicates that the "search.cgi" file cannot be executed for some reason, please check with your web host regarding your web server configuration.

It is also likely that if your "infomed_search.php" script is attempting to execute the "search.cgi" file, you will have to verify if PHP (or Apache) has permission / proper configuration to execute the CGI when it is running under the SSL configuration. Again, this can be checked by your web server administrator or web hosting company.

[phil@infomed.org]

Thank you Ray, let me try to answer the concerns one by one:

1) You're right, I had not properly copied over the "advanced" tab when installing v7. And then I couldn't find the field right away. Thanks for pointing it out to me. I am currently re-indexing, hopefully that will solve the first problem.

2) When I add https redirect in htaccess the output is "gibberish" in all cases.

3) Indexing https works but the results are the same as when indexing http. Ie. when the htaccess redirect to https is on the search results page shows "gibberish".

4) Correct, I seem to be able to index w/ v7 even when htaccess is redirecting to https. (but I didn't go through with the entire test since the indexing the site takes about half an hour).

It seems that the problem now is just that the cgi-bin is not executing properly when the htaccess is on. Is this assumption accurate?

Thanks
ph

[Ray]

5) However, when trying to do a search from our actual search page (www.infomed.ch/infomed_search.php) >> the search fails, i.e. it's a file not found msg bc the link is wrong (you can test this now on the live site)
6) The same thing happens when I try to order the search results by date or when I try to call up a sub-category of search results

This looks like it's missing the "Link Back URL" setting (under "Configure"->"Advanced"). This should be set to "infomed_search.php" since you are using a wrapper page.

7) When I add the redirect to SSL in htaccess >> the search fails like before

Can you clarify if you mean it is reporting the "Could not download file... (Invalid URL or domain name)" error, or you mean the search page is failing like reported in step 5. The latter is explained above, and unrelated to the redirect.

I tried the same thing starting the index from https. The results were basically the same as above:
Searching from our header search window works, but doesn't work from the search page. Adding redirect in htaccess makes the search crash.

It's important to be consistent and clear with describing the problem. What are you describing as "makes the search crash"? Is it the "Could not download file..." error? Or is it the gibberish output? Or is it the "File not found" error? Technically, none of these are a "crash" (which is when an executable fails during execution and terminates prematurely). So this is ambiguous.

I also tested starting the indexing w/ the htaccess redirect on and that now seems to work. Ie. I did not index the entire site bc I didn't want to leave the redirect on for so long but the indexing seems to work, it does not crash like it did previously in v6.

OK, so this means V7 is indexing better, and you're no longer getting the "Could not download file..." error? If that's the case, we can narrow down what we're looking at.

>> What is to be done about the problems w/ the redirect in htaccess, does this have to do with a setting that changed when SSL was added?

It's not clear which problem is persisting, that is related to the redirect. Are you still getting the gibberish output?

[phil@infomed.org]

Hi again,

Here are some results w/ the new versions of Zoom:

1) Installed v7 on a newer windows machine. No problems installing. Manually copied the configuration for infomed.ch over to the new install.

2) Indexed the complete site from http:// w/ no redirect in htaccess >> all good
3) Uploaded files manually via ftp (trying not to break our search pages)
4) Did a search from the search window in the header of our pages >> all good, search results show up

5) However, when trying to do a search from our actual search page (www.infomed.ch/infomed_search.php) >> the search fails, i.e. it's a file not found msg bc the link is wrong (you can test this now on the live site)
6) The same thing happens when I try to order the search results by date or when I try to call up a sub-category of search results

7) When I add the redirect to SSL in htaccess >> the search fails like before


I tried the same thing starting the index from https. The results were basically the same as above:
Searching from our header search window works, but doesn't work from the search page. Adding redirect in htaccess makes the search crash.

I also tested starting the indexing w/ the htaccess redirect on and that now seems to work. Ie. I did not index the entire site bc I didn't want to leave the redirect on for so long but the indexing seems to work, it does not crash like it did previously in v6.

So I have a couple of questions:

>> Why does the search work from the search window in the header but not from the search page?
>> How do I fix the links and the search interface in our search page so that the we don't end up on an empty page. Is this something we need to adjust on our search page?

>> What is to be done about the problems w/ the redirect in htaccess, does this have to do with a setting that changed when SSL was added?
>> Is there something I need to ask our hosting company to do?

Thanks
ph

[David]

The procedure entry point GetTickCount64 could not be located

Yes, this an XP issue. This function was only implemented in Vista and above.
In Version 7.0.1024 (7 Oct 2015) we dropped support for Windows XP.

[phil@infomed.org]

Question ...

I've installed v7 of the indexer on my PC but and tried to copy over the configuration for infomed.ch, however, when I try to start indexing I get a message reading "The procedure entry point GetTickCount64 could not be located in the dynamic link library KERNEL32.dll". And no indexing happens.

Is this due to the fact that I have the indexer installed on a Windows XP machine?

Best
ph

[phil@infomed.org]

Thank you Ray for your detailed response.

>> Currently there is no redirect in the htaccess file, but SSL is on.
>> Yes, we are using CGI
>> I wasn't aware that we needed to re-assign permissions w/ the SSL certificate in place. Will look into that.
>> We have not changed anything on our firewall but will check if that is an issue once we do the upgrade of the indexer

>> I will go through all the steps and upgrades and report back.

Thanks for your helo!
ph

[Ray]

For clarity, there's several different problems reported here. I'll try to summarize everything more clearly below.

Problem 1:
You are failing to index the website from a https:// URL. Zoom reports a "Could not download file... (Invalid URL or domain name)" error on the start https:// URL. You determined that this only happens when you have redirect enabled for all http:// URLs to https:// URLs.

Answer 1:

• You are using a very old build V6.0 build 1019 that is 8 years old. There is no recorded problem with HTTPS or redirects with this build or subsequent V6 releases, but there is at least 3 years worth of bug fixes for V6 that you are behind.
• V7 utilizes a completely new HTTP and HTTPS engine. So it would behave completely differently to V6.
• We have tested both the http:// URL and the https:// URL to work fine with the latest build at moment of writing (V7.1 build 1012) on your website. I tested it again just now and it appears that while the site is redirecting in the browser, Zoom still proceeds to index from either URLs fine and without issue.
• V6 is no longer supported, and we can not test the various old builds to verify what was or wasn't a problem in old discontinued releases.
• We also suspect it is likely a firewall issue, which you should investigate (e.g. temporarily disabling the firewall settings, or attempting to index from another computer)

Problem 2:
Your search results render as gibberish when redirect to SSL is enabled. [Shown in screenshot attached to your most recent post.]

Answer 2:

• We can see in the screenshot that, first of all, you have a PHP page (infomed_search.php). I presume this is executing a CGI, capturing the output and displaying it on the page (presumably like the wrappers described in our support page here. Let us know which method is being used, and if otherwise).
• There is quite possibly complications with redirections to SSL and whether PHP is set up to execute a CGI when running under SSL.
• The "gibberish" actually looks like a binary dump of the 'search.cgi' file. This means it is failing to execute the CGI and your PHP code is causing it to simply dump the contents of the CGI file.
• Interestingly, going to www.infomed.ch/cgi-bin/search.cgi causes the CGI file to download (with either http or https). This means it isn't setup to be executed properly. However, I can't confirm if this is the actual CGI file you are calling from infomed_search.php, or a stray copy from a previous setup attempt.

Given the above, it is possible the first problem is either Firewall related, or can be fixed by switching to V7. We cannot guarantee this without being able to confirm if the problem is your firewall or PC.

The second problem is very unlikely to be fixed by upgrading to V7. It is almost definitely a web server configuration issue.

Having said all that, as we no longer support V6, I can't continue to investigate this issue unless it is a problem with V7. So I hope the above summary of where we're at has cleared up the case and will help you get to the bottom of it.

[phil@infomed.org]

Hi again,

Thanks for your help.
The search currently works bc we removed the redirect in htaccess.
Here you will find a screen shot w/ the SSL redirection in htaccess enabled: https://dl.dropboxusercontent.com/u/...ss_problem.pdf

But understood about the release version.
I don't mind doing the upgrade but I want to make sure that this will fix the problem. Do you have any record about this being an issue in the past and it being fixed in newer releases?

Best
ph

[Ray]

Just noticed you're actually using V6.0 build 1019. That is a very old release (8 years) and no longer supported. If there was a known bug in V6 it may have been fixed in the final V6.0 release available from here. It is also very possible Windows has changed since then and it doesn't support it.

We would advise updating to the latest version.

Upgrade details are here.

[Ray]

We've verified that there's no problems indexing from a https:// URL with the latest build (V7.1 build 1012).

I've also just tried indexing from your actual site, with either of the following as Start URLs
https://www.infomed.ch/index.php
http://www.infomed.ch/index.php

Both indexed fine. I didn't see any redirect in either cases.

So if you are still experiencing a problem, please check the firewall for your indexing computer.

[phil@infomed.org]

Hi again,

Thanks for replying.
When the redirect is on nothing works ...
>> I tried indexing from http
>> indexing from https
Both fail.

Also, the search page on the website doesn't work.
But without the redirect the search works, even on https.

Best
ph

[David]

When you index is you start point the HTTP address or the HTTPS address?
If you hit the redirect, then I assume you are starting at the HTTP address. Try replacing the start point with the HTTPS address.

[phil@infomed.org]

Hi again,

So we've narrowed the problem down to the htaccess file. As soon as we add a redirect in htaccess to send all traffic to https, the zoom search does not work anymore on the website and indexing is no longer possible from the windows machine.

Is there something specific in htaccess that could make the search engine fail?

Thanks
ph