Workflow not working: SuiteCRM installed on CentOS-6.3 64bit

jchrisj · 3 May 2017 10:22

Before I jump into actions according to your instruction. I would like to express my deep gratitude to you for your great help. And I would like to tell you that I love the first line most because I can feel your deep passion in helping others as well as in your profession. You deserve what your signature says “SuiteCRM Is My Life.” And I feel extremely happy and lucky getting you to help me.

I will following your instructions and will let you know the results.

Again, millions of thanks.

Chris

pgr:

You’re doing things without thinking, without understanding, and without reading what we write.

You were told to insert a line into the system-wide crontab which is the file at /etc/crontab, and it has a specific syntax with an extra column for the user name.

But you went and put that line into a different crontab, at /var/spool/cron/root, which is obviously not a system-wide crontab, it’s root’s crontab (the one we DON’T want), and has a syntax that does not expect a username.

So now you have a job running as root, and it’s using a command “nginx cd /a/domain…” etc, so the username is being interpreted as the first word of the command. That can never work.

So remove those two lines (why two?) from /var/spool/cron/root, add a single line with the username to /etc/crontab (that’s a file you can edit).

And don’t touch workflows before you get this right. : - )

jchrisj · 3 May 2017 10:59

I did what you instructed and still got no luck.

below are the screens photos:

change of the /etc/crontab:

crontab running:

process workflow tasks: still no results

It seems it still does not work. What is wrong?

Please instruct.

Thanks!!!

amariussi · 3 May 2017 11:21

I think that you entered a wrong path in the command.

Instead of:
/a/domains/lms.softpowerworks.com/public_html

Please try:
/a/domains/lms.softpowerworks.com

(I went to both these paths and the first, which is the one you used, does not seem to exist)
I also executed http://lms.softpowerworks.com/cron.php which seems to have worked fine.

jchrisj · 3 May 2017 12:05

still no luck

I got the following:

cron job running:

job log of “PROCESS WORKFLOW TASKS” still empty:

God. What is really going wrong???

Please help.

Thanks.

pgr · 3 May 2017 13:48

Yes.

If you enter these two commands

ls -al /a/domains/lms.softpowerworks.com
ls -al /a/domains/lms.softpowerworks.com/public_html

what results do you get? I’d like to see a cron.php file there, with the correct ownership and permissions.

If none of those directories contains a cron.php file, then look for one with

find  /   -name cron.php 2>/dev/null

jchrisj · 3 May 2017 13:49

I kept trying but get no luck.

Can you please continue to help me out?

Thanks.

pgr:

You’re doing things without thinking, without understanding, and without reading what we write.

You were told to insert a line into the system-wide crontab which is the file at /etc/crontab, and it has a specific syntax with an extra column for the user name.

But you went and put that line into a different crontab, at /var/spool/cron/root, which is obviously not a system-wide crontab, it’s root’s crontab (the one we DON’T want), and has a syntax that does not expect a username.

So now you have a job running as root, and it’s using a command “nginx cd /a/domain…” etc, so the username is being interpreted as the first word of the command. That can never work.

So remove those two lines (why two?) from /var/spool/cron/root, add a single line with the username to /etc/crontab (that’s a file you can edit).

And don’t touch workflows before you get this right. : - )

jchrisj · 3 May 2017 14:02

for ls -al /a/domains/lms.softpowerworks.com I got:

for ls -al /a/domains/lms.softpowerworks.com/public_html I got:

I can see cron.php under the folder /a/domains/lms.softpowerworks.com/public_html through ftp.

Do I need to make any changes in the cron.php file?

Thanks.

pgr:

Yes.

If you enter these two commands
ls -al /a/domains/lms.softpowerworks.com
ls -al /a/domains/lms.softpowerworks.com/public_html
what results do you get? I’d like to see a cron.php file there, with the correct ownership and permissions.

If none of those directories contains a cron.php file, then look for one with
find  /   -name cron.php 2>/dev/null

pgr · 3 May 2017 14:14

I don’t understand your permissions scheme. You have some files owned by user “www”, group “www”, others by user “nginx”, group “nginx”.

Depending on exactly how these groups are constituted, your web server (which we’re guessing operates as “nginx”, but we’re not really sure) might not be getting access to the file.

Let’s try resetting your permissions.


cd /a/domains/lms.softpowerworks.com/public_html
sudo chown -R nginx:nginx . 
sudo chmod -R 755 .
sudo chmod -R 775 cache custom modules themes data upload config_override.php

Note that some of those commands end with a space followed by a dot (.)

Also, the correct line in /etc/crontab should reference the /a/domains/lms.softpowerworks.com/public_html directory, so if you had changed that, please change it back.

If that doesn’t work (it might even break your entire application), try setting things up for the other user:


sudo chown -R www:www .

then change the username in crontab from “nginx” to “www”.

jchrisj · 3 May 2017 14:33

I tried everything before

sudo chown -R www:www .

and nothing happened.

Then I applied the above command and the application really broke. I can not see “Scheduler” any more…

What shall I do next?

Thanks.

pgr:

I don’t understand your permissions scheme. You have some files owned by user “www”, group “www”, others by user “nginx”, group “nginx”.

Depending on exactly how these groups are constituted, your web server (which we’re guessing operates as “nginx”, but we’re not really sure) might not be getting access to the file.

Let’s try resetting your permissions.
cd /a/domains/lms.softpowerworks.com/public_html
sudo chown -R nginx:nginx . 
sudo chmod -R 755 .
sudo chmod -R 775 cache custom modules themes data upload config_override.php 
Note that some of those commands end with a space followed by a dot (.)

Also, the correct line in /etc/crontab should reference the /a/domains/lms.softpowerworks.com/public_html directory, so if you had changed that, please change it back.

If that doesn’t work (it might even break your entire application), try setting things up for the other user:
sudo chown -R www:www . 
then change the username in crontab from “nginx” to “www”.

jchrisj · 3 May 2017 14:38

The whole SuiteCRM application is down now… help please!

pgr · 3 May 2017 14:49

What did you do? Is it down with ownership for nginx, for www, or for both?

jchrisj · 3 May 2017 22:23

It is down with ownership for www.

It is down after I use the following command:

sudo chown -R www:www .

I guess this is to change the ownership all to www, right? So the whole SuiteCRM on my service can not be owned by www.

Please help.

Thanks.

pgr · 4 May 2017 08:45

Well, why did you change the ownership of all files to www? Wasn’t it working with nginx?

Did you read my post when I gave you those commands?

amariussi · 4 May 2017 08:52

When I tried to see your website I noticed that the crm was working from the root folder and not the folder called public_html. However the files are in this latter folder.
I am not the right person to help on this but I believe that there may be something in your php.ini or in .htaccess or elsewhere to cater for this. You should check if that is the cause of your problem.

jchrisj · 4 May 2017 09:11

I did strictly according to your post. I also read very carefully this line “If that doesn’t work (it might even break your entire application), try setting things up for the other user:” as it contains “it might even break your entire application”.

Because all the command before

sudo chown -R www:www .

did not work. I applied the above command and the system turned down.

Afterward I did the following:

sudo chown -R nginx:nginx .

And the system came back again.

It seems the whole application can be run under user “nginx”, not “www”

But even SuiteCRM works with “nginx” user but the workflow automation still not working.

What do I need to do next? Do I need to restore the whole system? If I need to change php.ini or anything else, can you please help me?

Thanks.

Chris

pgr · 4 May 2017 10:21

@amariussi his web server seems to be serving the files from public_html, so that

http://crm.softpowerworks.com/index.php

will actually serve

/a/domains/crm.softpowerworks.com/public_html/index.php

I think it’s just a misleading directory name…

pgr · 4 May 2017 10:28

@jchrisj I told you to

1.set ownership to nginx and try it.
2. then IF it didn’t work, change it to www.

What you did:

set ownership to nginx
change it to www
complain it’s not working
change it back to nginx
say it’s working (but not the workflows).

Anyway, i think we now have our ownerships and permissions right, which is a great achievement after 35 replies in this thread!

Now, I wonder what you currently have in /etc/crontab?
I wonder what the cron logs say about cron execution?
I wonder what suitecrm.log has related to Scheduler jobs?
I wonder what you see in the screen Admin/Schedulers, do jobs appear as executing correclty?

I do NOT wonder about whether your Workflows are working, that’s something that only becomes relevant later…

jchrisj · 4 May 2017 10:53

Thanks for your help.

I did not do the following steps one by one. What I did is:

set ownership to nginx and try it.

But I see nothing happened. Then I did the next step:

change ownership to www.

What I did wrong is not reporting to you before doing step 2.

Anyway thanks so much and I would like to bother you to move forward:

Now, I wonder what you currently have in /etc/crontab?

this is what I have:

I wonder what the cron logs say about cron execution?

this is what I got.

I wonder what suitecrm.log has related to Scheduler jobs?

in suitecrm.log I have the following:

[ERROR] convert: Conversion of 2017-05-03 00:00 from m/d/Y H:i to Y-m-d failed
Thu May 4 13:02:44 2017 [42637][6aed3cad-88d0-621b-81f4-58f5788d11b1][ERROR] convert: Conversion of 2017-05-03 00:00 from m/d/Y H:i to Y-m-d failed
Thu May 4 13:03:00 2017 [42638][6aed3cad-88d0-621b-81f4-58f5788d11b1][ERROR] convert: Conversion of 2017-05-03 00:00 from m/d/Y H:i to Y-m-d failed
Thu May 4 13:34:55 2017 [42637][6aed3cad-88d0-621b-81f4-58f5788d11b1][ERROR] convert: Conversion of 2017-05-03 00:00 from m/d/Y H:i to Y-m-d failed
Thu May 4 18:17:53 2017 [45409][6aed3cad-88d0-621b-81f4-58f5788d11b1][/i]

I wonder what you see in the screen Admin/Schedulers, do jobs appear as executing correclty?

No. nothing happened.

Thanks again for your great patience.

jchrisj · 5 May 2017 14:37

@pgr

After very hard learning, I finally got the following message. can you please help me to get this resolved? Many thanks.

Fri, 05 May 2017 22:33:01 +0800
Date: Fri, 05 May 2017 22:33:01 +0800
Message-Id: E1d6eI1-000MtS-AY@iZ25pqi27dnZ
From: root@iZ25pqi27dnZ (Cron Daemon)
To: root@iZ25pqi27dnZ
Subject: Cron nginx@iZ25pqi27dnZ cd /a/domains/car.softpowerworks.com/public_html; php -f cron.php > /dev/null
Content-Type: text/plain; charset=UTF-8
Auto-Submitted: auto-generated
X-Cron-Env: <LANG=en_US.UTF-8>
X-Cron-Env: <SHELL=/bin/bash>
X-Cron-Env: <PATH=/sbin:/bin:/usr/sbin:/usr/bin>
X-Cron-Env: <MAILTO=root>
X-Cron-Env: <HOME=/>
X-Cron-Env: <LOGNAME=nginx>
X-Cron-Env: <USER=nginx>
Sender: CronDaemon nginx@iZ25pqi27dnZ

/bin/bash: php: command not found

pgr · 5 May 2017 15:05

Hi,

it’s not finding your php executable.

In a shell, type

which php

to get the full path. Then change the line in cron to include the full path, for example, if it is “/usr/bin/php”, the full line would be


* * * * * nginx cd /a/domains/car.softpowerworks.com/public_html; /usr/bin/php -f cron.php > /dev/null