Blank pages, unable to "/debug=yes/"

Help
1

I’ve started to work on my “other” instance of librenms while attempting to recover from the dreaded “Error: Missing dependencies! Run the following command to fix:” event.
I applied the missing php php71w-mbstring.x86_64 package, have successfully run validate.php and daily.sh rebooted the server and cleared my client caches for good measure, but I’m still getting blank screens.

What bothers me most is that it looks like I should be able to get debug information by appending /debug=yes/ to pages but that doesn’t seem to have any effect.

librenms.log is full of successful polls.

The server is CentOS Linux release 7.5.1804 (Core)

Component Version
LibreNMS 1.40
DB Schema 251
PHP 7.1.17
MySQL 5.5.56-MariaDB
RRDTool 1.4.8
SNMP NET-SNMP 5.7.2

====================================

[OK] Composer Version: 1.6.5
[OK] Dependencies up-to-date.
[OK] Database connection successful
[OK] Database schema correct

I even added a memory statement to config.php ($config[‘php_memory_limit’] = 128;) but at this point I’m starting to do things that may not be warranted. Before I start kludging around too much, can I get some suggestions?

Thanks

2

I seem to remember i had something similar, worth trying this

remove your php71w-mbstring and install the previous version

yum install php70w-mbstring

then run

./scripts/composer_wrapper.php install --no-dev

If no success, see if any errors in your javascript console in your browser

3

Thanks for that, odd error though…

-bash-4.2$ ./scripts/composer_wrapper.php install --no-dev
/usr/bin/env: php: No such file or directory

4

hmm try reversing that, do you get the same error? remove 70w and install 71w and then run it

After you switched back to php71w-mbstring, also make sure this is installed the cli with yum install php71w-cli

then please post the output of yum list installed | grep php

5

Thanks Chas.

This is what I have installed (a match to the other, working instance that I have);

yum list installed | grep php
mod_php71w.x86_64                  7.1.17-1.w7                         @webtatic
php71w-cli.x86_64                  7.1.17-1.w7                         @webtatic
php71w-common.x86_64               7.1.17-1.w7                         @webtatic
php71w-gd.x86_64                   7.1.17-1.w7                         @webtatic
php71w-mbstring.x86_64             7.1.17-1.w7                         @webtatic
php71w-mcrypt.x86_64               7.1.17-1.w7                         @webtatic
php71w-mysql.x86_64                7.1.17-1.w7                         @webtatic
php71w-pdo.x86_64                  7.1.17-1.w7                         @webtatic
php71w-process.x86_64              7.1.17-1.w7                         @webtatic
php71w-snmp.x86_64                 7.1.17-1.w7                         @webtatic
php71w-xml.x86_64                  7.1.17-1.w7                         @webtatic

Other than that I had to re-enter the php timezone and it would seem to be good to go.

Validate is complaining about cron, but that won’t have anything to do with this problem. I ran the daily and still have blank pages. I must be missing something small.

6

I’m surprised that you got this error when running composer, but validate.php works:
./scripts/composer_wrapper.php install --no-dev
/usr/bin/env: php: No such file or directory

The error is very similar to this guy, so might be worth trying this solution
#https://stackoverflow.com/questions/43878731/composer-installed-but-get-usr-bin-env-php-no-such-file-or-directory
sudo ln -s /usr/bin/php71 /usr/bin/php

and re-running ./scripts/composer_wrapper.php install --no-dev

7

I just noticed your php-fpm is missing, try yum install php71w-fpm

Then re-check from the install guides you configured php-fpm e.g
https://docs.librenms.org/#Installation/Installation-CentOS-7-Nginx/
Under “Configure and Start PHP-FPM”

8

I understand what you’re recommending, but the other instance I have should be just like this and neither run nginx. I definitely appreciate your help but at this point it probably makes more sense just to migrate to a fresh instance. Is there a recommended procedure for migrating the existing collected data and host info to a different host?

9

Looks like this is the method.

I’ll probably spin up a new host but leave the other around to see if I actually can fix it.

10

How did it go @pipegrep did you manage to fix it?

11

We just took over a server room so that’s become an immediate priority. I’ll reply here when I have closure, thanks.