Several Problems at Fresh Install


My setup;

  • DigitalOcean Debian 8.2 (firewall disabled)
  • Latest Apache & PHP from official repos
  • Percona XtraDB Cluster (2 node)

I got few problems;

Diagnosis Tools stuck at 15%. Here is where it stuck;

Getting config.php...

Getting custom dir...

Getting phpinfo()

Getting...... Database info... Database dumps... Database schema

No error at suitecrm.log, apache, php logs about this.

Also I have cron job problem;

I’ve tried;

To add root’s crontab as;

*    *    *    *    *     cd /var/www/; php -c /etc/php5/apache2/php.ini -f cron.php > /dev/null 2>&1 
*    *    *    *    *     cd /var/www/; php -f cron.php > /dev/null 2>&1 
*    *    *    *    *     su -s /bin/bash -c 'cd /var/www/; php -f cron.php > /dev/null 2>&1 ' www-data

To add www-data crontab;

*    *    *    *    *     cd /var/www/; php -f cron.php > /dev/null 2>&1 
*    *    *    *    *     cd /var/www/; php -c /etc/php5/apache2/php.ini -f cron.php > /dev/null 2>&1 

and It throw as error constantly;

[1][FATAL] Job b9c751dd-5b7f-4a8a-cbfd-567950f93349 (Run Report Generation Scheduled Tasks) failed in CRON run]

I’ve checked my permission 3 times. User and Group added at config.php (as www-data). Mail settings are correct.

Also my e mail setting are completely correct and working. But sending a mail takes more than 1 minutes, about several minutes. By that time, CRM doesn’t response.

By the way my php.ini setting are OK. Execution time memory limit etc.

I found a bug report at SugarCRM about Diagnosis Tools;

I downgraded my OS to Debian 7 which has PHP 5.4. All my problems still exist.

Downgraded to Centos 6.7 which has PHP 5.3.

Diagnosis stuck problem solved.

Other problems continues.

turn error reporting off?

Alot of errors and hangups with php version. I know ive gotten Suitecrm to run fine on php 5.4.

I hope at some point the code will be made fully compatible to new php. But we will see. I sure as hell dont know how to do it.

Latest and greatest when it comes to stability is never the case, Sometimes running an older build with all the security updates is just better for something you rely on. You know it works and has gone through the troubles and been fixed already.

sorry, ment to be specific here, Have you tried turning display_errors=Off in php.ini? Or changing the error types to be displayed?

while display_errors=On if i run manually the cron command it throws;

PHP Notice: Undefined variable: reminders in /var/www/ on line 241

And still got the same error at suitecrm.log.

I can provide detailed logs (debug level).

Also reported Diagnostic Tool problem at github; here is the link;

For last; would you provide your crontab settings? Which user’s crontab, what is the command itself etc…

This community is so dead.

Horrible, uber outdated documantation as classic. Otherwise how are they gonna earn money right? (sarcasm)

I won’t surprise if Sales Agility change it to closed source in short future.

Dude, its the holidays. People are busy with their family. And I honestly do not foresee this being put closed source any time. There is the whole clause of them legally having to include powered by sugarcrm on the bottom. You are talking about a small team of people developing and giving this away for free and you would like instant free support from a very small niche community.

Honestly my first start would be to go tried and true, that is run on centos6 even centos7 is fine as I run it on there as well. I’m confused I I just installed the latest non beta build of Suitecrm and had no issues.

So lets break everything down to one issue at a time. You have 3. I think the most important is your install, you said it was stuck, and then you mentioned cron and email not working. I would suggest getting your install to go through first, probably could be a source of all your issues since you cannot even complete an install.

As you notice the install according to you stops on the DB steps. You are not using a standard DB in which Sugar/Suite was intended to be used with. That is opensource mysql and now should be mariadb. From my limited understanding not all SQL DB’s are the same as they take different commands and shortucts etc. This could be an issue. I do not understnad why you are using Percona, when a perfectly fine opensource DB can be used that will handle your workload sufficiently. I have an office here of 40+ employees using a single shared server with a single DB on the same server in VPS handling complicated hooks and saves repeatedly non stop everyday, 8 hours a day. I rarely get peformance issues.

Try building a stack from scratch on Centos7.

Now I just realized you said install problem was fixed. OK so i’m fairly certain email is gone through batches from the cron script. I see cron error you get is specific to report generation. Try finding the offending report. There are no reports turned on by default that I am aware of so this is a report you have instituted.

AS far as email not going through, what service are yo uusing for smtp? your own? thirdparty? try to break down the steps the system goes through to send or receive an email and see where it breaks.

If you are running your own email stack, chances are that is the problem as its very difficult to have emails go through with your own stack from experience due to strict spam policies from all major email companies.

MY suggestion is to use a minimal OS, manually install the stack and php modules with mariadb. Correctly add the cron script to correct user. Use a supported DB server. How many clients do you foresee using your system that you believe you need Perconda XtraDB cluster?

And as far as support, i seriously think Ive only seen 2 or 3 official people respond, thats a very small amount of people. You cannot expect them to jump and guide you through an install right before a major Holiday. If you look at my post I only post every couple of months due to me being so busy. The only way for the community to grow is through people contributing the little they can. That includes yourself.

You are totally right about holiday. That was my bad.

I’ve tried to build stack (from scratch) at Debian 8, Debian 7, CentOS 7 and recently on CentOS 6.

For databases; I’ve faced horrible SQL scenarios many times. I will, never ever, use a mysql database without proper HA & Redundancy especially at mission critical situations like this. Percona uses stardart mysql client and there is no difference at command sets. Besides there is no clue about a database problem.

For “Run Report Generation Scheduled Tasks failed in CRON run” problem; Appeared out that it’s (probably) a known bug:

Also CRON Jobs crashes if you create a report (any report let’s say lead source-referred by report). As told before I would be delighted to provide all my logs to dev if they care of course.

E-mailing is at third party. Perfectly working IMAP and SMTP servers (with SSL support). Very slow emailing problem continues. I’m spending my time on it. Will inform here if I find something.

As I report, Diagnostic Tool problem is most probably about the same problem with SugarCRM’s relevant bug. No problem on CentOS 6, PHP 5.3 (not even a warning at the logs)

About “support”; I don’t want to get “premium” support by “jumped-in” staff. In fact I always prefer to solve my problems on my own with PROPER DOCUMENTATION which is BACKBONE of open source projects. Good documentation will lead good community. We don’t even have correct “prerequisites” list.

Found the problem. IPv6 nameresolving takes forever. It might my setup problem instead of general SuiteCRM problem. Disabling v6 solved my problem.

I’ve focused to “Run Report Generation Scheduled Tasks failed in CRON run” problem now.