ERP5

We tryng to install https://www.erp5.com/deploy in our company, but no succeded. We have ubuntu 

Distributor ID: Ubuntu
Description:    Ubuntu 16.04.6 LTS
Release:        16.04
Codename:       xenial
 

At the 1st step we got the same https://www.erp5.com/forum/ERP5-wont-install-on-a-Linux-VM-FbAx5kNAac error ( u'ffff:ffff:ffff:ffff::/64' is not a valid IPv6 address string! ) . But is was gone with ubuntu upgrade. New we got the following error:

Traceback (most recent call last):
  File "/opt/slapos/eggs/slapos.core-1.5.4-py2.7.egg/slapos/grid/slapgrid.py", line 1335, in processComputerPartitionList
    self.processComputerPartition(computer_partition)
  File "/opt/slapos/eggs/slapos.core-1.5.4-py2.7.egg/slapos/grid/slapgrid.py", line 1226, in processComputerPartition
    raise e
IOError: Software Release https://lab.node.vifib.com/nexedi/slapos/raw/1.0.112/software/erp5/software.cfg is not correctly installed.
Missing file: /opt/slapgrid/5b41896bce906eba84123c9581f4751d/instance.cfg
 

Is there software.cfg ok (I see fail message)?

[buildout]
extends =
  ../../stack/erp5/buildout.cfg
# Test Suite: ERP5-MASTER ran at 2019/06/20 17:16:16.546716 UTC
#             10 failures, 0 errors, 6038 total, status: FAIL
[erp5]
revision = 80994259537cc30d255d9230aabaf47dce621bf8

How we can fix this issue?

I think it was a network problem. If you try again, it should work.

From: Unknown User

From: Yusei Tahara

I think it was a network problem. If you try again, it should work.

 

I think it's related to https://github.com/Nexedi/erp5/commit/b5b16628d0753647b87b4b7ff6c72eb32c3b224f#diff-8903239df476d7401cf9e76af0252622  .

You remove submodule "doc-bt5", but installer don't know about it. 

 

Additional info:

/opt/slapos/log/slapos-node-software.log 

[2019-12-21 14:55:43,797] INFO     Warning: "branch" parameter with value "master" is ignored. Checking out to revision 80994259537cc30d255d9230aabaf47dce621bf8.

So, it clone https://lab.nexedi.com/nexedi/erp5/tree/80994259537cc30d255d9230aabaf47dce621bf8 revision , not master.

/opt/slapos.playbook/erp5-standalone.yml

vars:
    software_release_url: https://lab.node.vifib.com/nexedi/slapos/raw/1.0.112/software/erp5/software.cfg

Is that correct version (1.0.112) of  a release?

Hi. I got new information.

I am sorry but latest version of ERP5 is currently not building.

Our build test of ERP5 into a Debian Stretch OS is currently failing :

You can track all our tests on this page:

https://stack.nexedi.com/test_status?fbclid=IwAR0tnOIWOPn3ksRLnQES06zBe5yYNBffK3c8FwB502BGn2VG1yFZr_5nCdw

We are now fixing the build test of ERP5 on Debian. Until it is fixed, ERP5 will not build.

From: Unknown User

Hi. I got new information.

I am sorry but latest version of ERP5 is currently not building.

Our build test of ERP5 into a Debian Stretch OS is currently failing :

You can track all our tests on this page:

https://stack.nexedi.com/test_status?fbclid=IwAR0tnOIWOPn3ksRLnQES06zBe5yYNBffK3c8FwB502BGn2VG1yFZr_5nCdw

We are now fixing the build test of ERP5 on Debian. Until it is fixed, ERP5 will not build.

 

Ok, I've got it, thanks for the info! It seems we will move to another erp system.

1. Any estimate as to when ERP5 build will be functional again?
2. Will this functionality only be for Debian 9 (Stretch), or will it also include Debian 10 (Buster)?
3. What seems to be the root cause of these failures and are these failures equivalent in both Debian 9 and 10 (according to the logs, script output, file presence, and computer behavior from our own testing, it seems to be the *.ymls in the playbook)?
I can possibly help fix the build, more specifically, the Debian 10 build, since we are attempting to run ERP5 on a brand new server Running Debian 10.2.0, however I will need to know where to start looking if I do so. I will keep an eye out on this forum post for responses. All the best!

I confirmed that the current ERP5 software release in master branch is buildable on debian 10. So, once the installation script is fixed, ERP5 will be installable on debian 10. It means that developers who know slapos can build erp5 on debian without installation script now.

From: Thomas Gambier

Hello, I'm responsible for upgrading the installer so that it works on Debian 10.

 

Yesterday, I fixed the problem you had in apache: https://lab.nexedi.com/nexedi/slapos/commit/e018d9df44243cf35340db0b0d1b88eafaa2e0a3

 

Today, I released a new version and push it into the playbook so you can already try the new version (see my latests commits in the installer script: https://lab.nexedi.com/nexedi/slapos.package/commits/master)

 

 I'm currently building everything and pushing it into our software cache. Once this is done, the test should work again.

 

Sorry for the inconvenience, we started to port everything to Debian 10 quite lately.

 

 

Thanks for the quick reply.

Funny you should mention Apache. According to a run I did on VirtualBox running Debain 9 about a week ago, the install script/playbook is responsible for installing Apache2 to provide the frontend to ERP5. I just ran the install script again yesterday around 3PM EST, but this time for the actual Debian 10.2.0 server we plan on using, and then ran the following today:
root@erp:~# systemctl status apache2

Unit apache2.service could not be found.

This means that the the install script never reaches the Apache installation instructions in the install script (which I believe to be somewhere in the playbook).

According to the slapos-node-software.log file (linked to mega due to size), the playbook does complete, but a quick search of the word error in gedit shows that the installation experiences several GCC errors, especially toward the beginning of when the build starts doing things in the background.

You did mention that you are able to install ERP5 manually by not using the script. Is there any instruction out there on how to do so? If not, I would find it beneficial to have one out there in-case the script breaks like this again, especially for new potential developers such as myself.

Although the server we are testing is meant to be used for a professional business, we are in no rush to get it up and running at the moment and are willing to lend it to be temporarily used as a testing platform for fixing the ERP5 installation on the Debian 10 distro of ERP5. The server is a brand new Dell PowerEdge T440 with 8GB RDIMM RAM and a Xeon 4210 Silver Processor. In addition, you’ll get the support of a Software Engineer as well as a Computer Engineer. We would like to help in any way we can as we see ERP5 as a way to help out other small businesses like our own looking to find an alternative ERP software that does what we need and isn’t ridiculously expensive.

I await your response and would like to know how we can help! Thank you!

This is a rough quick instruction of how to setup a standalone erp5 without installation script. (https://slapos.nexedi.com/ for more details)

First, your machine needs ipv6 internet connection. Then

apt-get install gnupg
wget -O- "https://download.opensuse.org/repositories/home:/VIFIBnexedi/Debian_10/Release.key" | apt-key add -

apt-get install slapos-node git

export PATH=$PATH:/sbin

slapos configure local (setup slapos node as standalone, it means that this node does not become a part of slapos cloud,
instead it uses a local dummy slapos master node, slapproxy)
)

slapos node format --now (It takes some minutes)

git clone https://lab.nexedi.com/nexedi/slapos.git

slapos supply /YOUR-SLAPOS-REPOSITORY-PATH/slapos/software/erp5/software.cfg local_computer
(local_computer is your node name, look at /etc/opt/slapos/slapos.cfg)

tail -f /opt/slapos/log/slapos-node-software.log (It takes several hours to build ERP5)
(`slapos node software` is executed by cron /etc/cron.d/slapos-node)
/opt/slapgrid/XXXX/.complete is a mark of complete.

prepare a request script like this:

import json
software_url = '/home/a/slapos/software/erp5/software.cfg'
parameter_dict = {}

request(software_url,
  'my_test_erp5',
  filter_kw={
    'computer_guid': 'local_computer',
  },
  software_type='default',
  partition_parameter_kw = {
    '_': json.dumps(parameter_dict, indent=2)
  }
)

slapos console < request.py

Let's wait for soem minutes. ERP5 cluster is setup automatically.
(`slapos node instance` is executed by cron /etc/cron.d/slapos-node)

Once setup is done, then,

slapos node

you will find something like this:

slappart0:bootstrap-monitor                                                 EXITED    Jan 31 10:31 AM
slappart0:caucased-on-watch                                                 RUNNING   pid 1134, uptime 0:00:07
slappart0:certificate_authority-1a4cd56cd8e659d9137c7bbd3db35a4e-on-watch   RUNNING   pid 1133, uptime 0:00:07
slappart0:crond-1a4cd56cd8e659d9137c7bbd3db35a4e-on-watch                   RUNNING   pid 1130, uptime 0:00:07
slappart0:monitor-httpd-1a4cd56cd8e659d9137c7bbd3db35a4e-on-watch           RUNNING   pid 1131, uptime 0:00:07
slappart0:monitor-httpd-graceful                                            EXITED    Jan 31 10:31 AM
slappart1:bootstrap-monitor                                                 EXITED    Jan 31 10:31 AM
slappart1:certificate_authority-1a4cd56cd8e659d9137c7bbd3db35a4e            RUNNING   pid 1138, uptime 0:00:07
slappart1:crond-1a4cd56cd8e659d9137c7bbd3db35a4e-on-watch                   RUNNING   pid 1141, uptime 0:00:07
...
slappart6:apache                                                            RUNNING   pid 1094, uptime 0:00:07
...

then, check your apache conf. In my case, it was in slappart6.

less /srv/slapgrid/slappart6/etc/apache/apache.conf

You will find which IP and port apache is listening. Apache uses internal IP,
thus it is not accessble from outside. You need to setup your own frontend web server to publish
slapos's apache. You should not change slapos apache conf by hand because it is overwritten by
"slapos node instance".

Once you can access to your erp5 instance, check your zope password.

less /opt/slapos/slapproxy.db (find inituser-password)

https://YOURERP5/erp5/ (use zope password to login)

From: Yusei Tahara

This is a rough quick instruction of how to setup a standalone erp5 without installation script. (https://slapos.nexedi.com/ for more details)

First, your machine needs ipv6 internet connection. Then

apt-get install gnupg
wget -O- "https://download.opensuse.org/repositories/home:/VIFIBnexedi/Debian_10/Release.key" | apt-key add -

apt-get install slapos-node git

export PATH=$PATH:/sbin

slapos configure local (setup slapos node as standalone, it means that this node does not become a part of slapos cloud,
instead it uses a local dummy slapos master node, slapproxy)
)

slapos node format --now (It takes some minutes)

git clone https://lab.nexedi.com/nexedi/slapos.git

slapos supply /YOUR-SLAPOS-REPOSITORY-PATH/slapos/software/erp5/software.cfg local_computer
(local_computer is your node name, look at /etc/opt/slapos/slapos.cfg)

tail -f /opt/slapos/log/slapos-node-software.log (It takes several hours to build ERP5)
(`slapos node software` is executed by cron /etc/cron.d/slapos-node)
/opt/slapgrid/XXXX/.complete is a mark of complete.

prepare a request script like this:

import json
software_url = '/home/a/slapos/software/erp5/software.cfg'
parameter_dict = {}

request(software_url,
  'my_test_erp5',
  filter_kw={
    'computer_guid': 'local_computer',
  },
  software_type='default',
  partition_parameter_kw = {
    '_': json.dumps(parameter_dict, indent=2)
  }
)

slapos console < request.py

Let's wait for soem minutes. ERP5 cluster is setup automatically.
(`slapos node instance` is executed by cron /etc/cron.d/slapos-node)

Once setup is done, then,

slapos node

you will find something like this:

slappart0:bootstrap-monitor                                                 EXITED    Jan 31 10:31 AM
slappart0:caucased-on-watch                                                 RUNNING   pid 1134, uptime 0:00:07
slappart0:certificate_authority-1a4cd56cd8e659d9137c7bbd3db35a4e-on-watch   RUNNING   pid 1133, uptime 0:00:07
slappart0:crond-1a4cd56cd8e659d9137c7bbd3db35a4e-on-watch                   RUNNING   pid 1130, uptime 0:00:07
slappart0:monitor-httpd-1a4cd56cd8e659d9137c7bbd3db35a4e-on-watch           RUNNING   pid 1131, uptime 0:00:07
slappart0:monitor-httpd-graceful                                            EXITED    Jan 31 10:31 AM
slappart1:bootstrap-monitor                                                 EXITED    Jan 31 10:31 AM
slappart1:certificate_authority-1a4cd56cd8e659d9137c7bbd3db35a4e            RUNNING   pid 1138, uptime 0:00:07
slappart1:crond-1a4cd56cd8e659d9137c7bbd3db35a4e-on-watch                   RUNNING   pid 1141, uptime 0:00:07
...
slappart6:apache                                                            RUNNING   pid 1094, uptime 0:00:07
...

then, check your apache conf. In my case, it was in slappart6.

less /srv/slapgrid/slappart6/etc/apache/apache.conf

You will find which IP and port apache is listening. Apache uses internal IP,
thus it is not accessble from outside. You need to setup your own frontend web server to publish
slapos's apache. You should not change slapos apache conf by hand because it is overwritten by
"slapos node instance".

Once you can access to your erp5 instance, check your zope password.

less /opt/slapos/slapproxy.db (find inituser-password)

https://YOURERP5/erp5/ (use zope password to login)

 

 

Glad you posted those instructions. No wonder there's an install script. That being said, I would recomend that ERP5 post that somewhere so that if the script fails again, developers can use it as a reference to figure out exactly which step the installation failed.
It also helped me pinpoint the issue that the install script is currently having with Debian 10. Based on everything you guys have told me and what I've found: using yesterday's build, I found that ERP5, zope, SlapOS and the apache backend are all present ( I found the apache.conf in partition 11; I also verified that the zope password already exists). This means that either the bash script or playbook completely fails at setting up the "frontend webserver to publish slapos's apache". I can't clearly tell if the frontend setup is meant to be in the bash script or in the playbook, but based on a sweep of both, neither contains instructions for setting up an apache frontend, or any frontend at all.

I'm not sure I fully understand how to "publish slapos's apache" from existing backend to frontend (I've only ever programmed frontend-only apache websites), but I'm also willing to wait for the script fix if you guys think you can fix the frontend installation problem in a reasonable time. As stated before, we're willing to help test the install script for ERP5 until its functional on Debian 10.

Thanks again!