• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Please help: Horde/String.php issues after PHP 5.2 Upgrade

Scott from atomic has a much cleaner solution.



Note this file is changed every time plesk is updated, so you'll need to go back and fix this when it does.

Is there a way to get this to auto fix? I have to update this thing every day. Its getting to be a annoyance
 
adamluz,

Are you actually changing things on your server causing the zz010_psa_httpd.conf file to be changed?

Now if not (make sure its NOT) how about this simple trick!

Log in as root, and do this:

cp /etc/httpd/conf.d/zz010_psa_httpd.conf /etc/httpd/conf.d/zz010_psa_httpd.conf.copy

(this copies your modified and working zz010_psa_httpd.conf and makes a copy zz010_psa_httpd.conf.copy)

Then set up a root cron, I would think hourly is enough, but if you know what is causing the issue adjust the time period) the cron MUST be root.


cp /etc/httpd/conf.d/zz010_psa_httpd.conf.copy /etc/httpd/conf.d/zz010_psa_httpd.conf

/etc/init.d/httpd restart

Now your backup copy rewrites over the messed up copy and restarts httpd
 
Thanks, thats a great idea. For anyone else trying this fix, Add --reply=yes to the cp line

cp --reply=yes /etc/httpd/conf.d/zz010_psa_httpd.conf.copy /etc/httpd/conf.d/zz010_psa_httpd.conf


This confirms to save over the original
 
however if you restore old copy, then it means that you loose all the further changes which are done in apache config by plesk -- no new domains, no modified hosting options..

instead i would prefer to copy webmail:80 and webmail:443 virtualhosts from conf.d/zz010_psa_httpd.conf and save them into conf.d/webmail.conf and correct webmail.conf instead.. as apache configs are loaded in alphabetical order, your custom virtualhost definitions in webmail.conf will be used and the next ones in zz010_psa_httpd.conf will be ignored..
 
Fixed

Hello,

ok this issue is fixed in Version 8.3 I updated my Etch without any problems and now I am using PHP 5.25 with Plesk 8.3 with correct working webmail *happy me* :)

Br,

dkuwi
 
Nice to see a professional to the point thread with direct solutions posted. Thought you all might like to know none of this is an issue for me. I'm running with Fedora Core 7, Plesk 8.2.1, PHP 5.2.5 and MySQL 5.0, and I have zero quarks of this nature.
 
I added the /lib mods - Still BLANK HORDE

I added the /lib mods and I am still getting blank horde pages....

Plesk 8.3 (was 8.1), tried 8.1, 8.2, 8.3.

SUSE 64 bit dual Xeon.

I posted some here about it....
http://forum.swsoft.com/showpost.php?p=195759&postcount=6

But I am not sure I am missing something but my problem is only showing up in the apache error logs as:

/usr/sbin/httpd2-prefork: symbol lookup error: /usr/lib64/php5/extensions/gettext.so: undefined symbol: php_realpath

Even after the mods.
 
This is really strange - the errors come back once in awhile, but as soon as I restart httpd (service httpd restart) everything works fine again, but only for a few minutes, then it reverts back to not working.

Anyone have any ideas?
 
Yea that is what I've been doing - but I thought 8.3 was supposed to fix this? Also, why is it that when I simply restart httpd it fixes it temporarily, then goes back to not working?
 
PSA cant really "fix" this, its a core behavior change in php. What they could "fix" was to add the required include path for horde.

I actually said why its doing this after a while in my post:

"From what Ive determined so far this is entirely based on the context of the first virtual host to invoke it..."

Ergo, if you are using domain-A with no php_admin_values, and someone else uses domain-B with php_admin_values (set in vhost.conf for example). Then it will effect both domain-A and domain-B, depending on when domain-B is invoked.
 
Here is a quick perl script I wrote to handle the problem of this file getting re-created with new information as domains are added/etc. It checks to see if the file needs fixing first and reloads apache after patching the file if necessary. I run it in cron on my plesk server that has this issue every hour.

-------------------------------------------

#!/usr/bin/perl

my $input_file = "/etc/httpd/conf.d/zz010_psa_httpd.conf";

my $tmp_file = $input_file . ".swp";

open CNF, $input_file or die "Could not open $input_file for read: $!\n";
open SWP, "> $tmp_file" or die "Could not open $tmp_file for write: $!\n";

while(<CNF>) {
$_ =~ s/php_admin_value include_path "\/usr\/share\/psa-horde:\/usr\/share\/psa-horde\/pear:."/php_admin_value include_path "\/usr\/share\/psa-horde\/lib:\/usr\/share\/psa-horde:\/usr\/share\/psa-horde\/pear:."/;
print SWP $_;
}
close(SWP);
close(CNF);

#check to see if the files are different
my $output = `diff $input_file $tmp_file`;

if($output ne "") {
print "Files differ. moving swap file to original filename...\n";
system("mv -f $tmp_file $input_file");
print "Reloading apache..\n";
system("/etc/init.d/httpd reload");
}
else {
print "Files are the same. Clearing temp file...\n";
system("rm -f $tmp_file");
}

print "Done\n";
 
Back
Top