Sudden crash of version 1.06

[skylark]

Everything was going great after our cmsms upgrade to 1.06 from 1.02.

Then suddenly after a couple of days, this error came up. I had just installed 2 modules, just before then...forum and pageblock. I don't know if the problem is related.

I have re-installed, and recopied the files to no avail....

The error when going to the homepage or the admin screen looks like this:

Warning: newdatadictionary(/home2/tobegro/public_html/tobe/cms/lib/adodb_lite/adodbSQL_drivers//_datadict.inc) [function.newdatadictionary]: failed to open stream: No such file or directory in /home2/tobegro/public_html/tobe/cms/lib/adodb_lite/adodb.inc.php on line 158

Warning: newdatadictionary() [function.include]: Failed opening '/home2/tobegro/public_html/tobe/cms/lib/adodb_lite/adodbSQL_drivers//_datadict.inc' for inclusion (include_path='.:/usr/lib/php:/usr/local/lib/php') in /home2/tobegro/public_html/tobe/cms/lib/adodb_lite/adodb.inc.php on line 158

Fatal error: Cannot instantiate non-existent class: adodb2_ in /home2/tobegro/public_html/tobe/cms/lib/adodb_lite/adodb.inc.php on line 161

Checked with the host, and they said that they made no changes to sql or the directory structures.

[tsw]

check all paths in config.php and upload all files again just to be sure

[skylark]

I did that.

Steps taken so far:

Checked out and verified with the host that the root path was valid.
Did a complete re-install of  cmsms version 1.06. Kept the database intact.

Exported and imported the mysql database etc.

The records and database look fine as far as I can see.

Unfortunately, the host's mysql database dump wasn't working and had submitted a trouble ticket to them at the time the day before. I have a backup of the database but it has garbage chararcters in it. The backup I have today, does not.

I deleted the records in the module table that I had just added before encountering the trouble: forum and pageblock.

This had no affect.

Now, I am looking at the adodb2...does anyone know that that is? I changed it in the config file from "true" from "false" but it had no affect.

Debug flag to "true" doesn't produce any more informational messages that I can see.

I am just now contacting the host about the adodb2 module and see if  maybe ithere could be memory problems since it can't instantiate or if it is physically where it should be in the directory structure. Since everything was working fine until 2 days ago, then I suspect memory problems or the host moving the adodb2 on us...

Anyone have any other ideas? We are totally down.

[calguy1000]

have you checked your httpd error log? maybe there's a memory limit problem or something minor like that.  or a permissions error causing an error in the script execution.

[skylark]

Thank you, yes I checked httpd error logs, and everything seems ok. I couldn't see anything that jumped out at me, anyway.

The host support had this to say:

>>It appears that the installation process had failed to
>>properly finish or the installation had become damaged after
>>the initial setup. The reason I say this, is because the
>>CMS script is not including the \'mysql\' line in the string
>>it uses to locate the datadict.inc and other scripts. Files
>>do, in-fact, exist that are named mysql_datadict.inc but the
>>CMS installation is not properly looking for them.

He (or she) suggested that I do a re-install without writing over the tables. I did that.
It didn't solve the problem.

Where do we define the pathnames to the datadictionaries...? I didn't find it in the config.php.
Is there another file we can configure the pathnames?

[tsw]

check that $config['dbms'] = 'mysql';

if that doesnt help

[skylark]

We seem to be on the right track in that the 'dbms' is coming up blank. But, the value #config[dbms']='mysql' is in the config file.

Is the dbms defined anywhere else???

How can it lose its value?

This is so frustrating, we have been down over a WEEK!
I think at this point, I will try re-installing back versions...

[tsw]

#config[dbms']='mysql'

with # character in front??

take # of from that line

[skylark]

There are no pound signs in front...double checked..I don't know why I typed it in there like that.

I moved everything to another host...same thing.

So, now at least we know it's not a host environment problem, or it's highly unlikely.

Any other ideas...it seems that our host support is on the right track in that the dbms database doesn't seem to be set when the string is created...

Any ideas?