Currently I am playing around with Bahmni as we consider using it for a charitable hospital in the Punjab. While I was investigating the patient merge feature of OpenMRS I had to try out to update Bahmni. I ran into problems and report it in case one of your real clients run into this issue as well.
Observed result
Ansible task “Remove maintenace enable page from /var/www/html” fails because it cannot access port 80 during update to Bahmni version 0.86. This could be reproduced for the update from 0.83 and 0.85.
{
# Workaround mentioned in https://bahmni.atlassian.net/wiki/display/BAH/Install+Bahmni+on+CentOS
export postgres_repo_rpm_name="pgdg-centos92-9.2-7.noarch.rpm"
# Setup instructions from https://talk.openmrs.org/t/how-to-update-bahmni-to-latest-version/7731
yum remove bahmni-installer -y
yum install https://dl.bintray.com/bahmni/rpm/rpms/bahmni-installer-0.86-66.noarch.rpm -y
bahmni -i local install
}
Environment
CentOS 6.8 on VirtualBox (without Vagrant)
Host sytem: Ubuntu 16.04
Diagnostic information
Problem analysis
The problem is that the Ansible jobs fail to release port 80 after stopping the httpd service. I have found this out by adding pauses into the following scripts and then I have logged into the VM with another terminal. The funny thing is that the httpd service is stopped but still port 80 is occupied.
[root@kch ~]# service httpd status
httpd is stopped
[root@kch ~]# nmap -sT -O localhost
Starting Nmap 5.51 ( http://nmap.org ) at 2016-11-28 03:09 IST
Nmap scan report for localhost (127.0.0.1)
Host is up (0.000061s latency).
Other addresses for localhost (not scanned): 127.0.0.1
Not shown: 994 closed ports
PORT STATE SERVICE
22/tcp open ssh
25/tcp open smtp
80/tcp open http
If I kill the process manually within the pauses indicated below with the following command, the installation will succeed.
This is awesome analysis. We faced this problem few times while testing fresh install of Bahmni 0.86. But, we couldn’t reproduce it. Which version of Ansible is installed on your system?
Hi @ajjaswanth,
I am kind of doing such things all day long. In case you are interested, here is the process I am trying to follow.
My Ansible version is 2.2.0.0. What might have caused the problem is that I am running an CentOS 6.8 while your pipeline seems to be running on CentOS 6.7. The problem with 6.8 is that it is not available for download on the official website. I think DigitalOcean also offers only 6.8/7.X.
The problem is not caused by the CentOS or Ansible version.
Experiments
Start of update with Ansible 2.2.0.0
Experiment
(1) Load snapshot 0.83
(2)
{
# Workaround mentioned in https://bahmni.atlassian.net/wiki/display/BAH/Install+Bahmni+on+CentOS
export postgres_repo_rpm_name="pgdg-centos92-9.2-7.noarch.rpm"
# Setup instructions from https://talk.openmrs.org/t/how-to-update-bahmni-to-latest-version/7731
yum remove bahmni-installer -y
yum install https://dl.bintray.com/bahmni/rpm/rpms/bahmni-installer-0.86-66.noarch.rpm -y
bahmni -i local -av 2.2.0.0 install
}
Result
Resulted in the same problem.
TASK [httpd-maintenance : reload_httpd_service] ********************************
task path: /opt/bahmni-installer/bahmni-playbooks/roles/httpd-maintenance/tasks/main.yml:23
failed: [localhost -> localhost] (item=localhost) => {"failed": true, "item": "localhost", "msg": "[Sun Dec 04 21:37:58 2016] [warn] module cache_module is already loaded, skipping\n[Sun Dec 04 21:37:58 2016] [warn] module disk_cache_module is already loaded, skipping\n[Sun Dec 04 21:37:58 2016] [warn] module wsgi_module is already loaded, skipping\nhttpd: apr_sockaddr_info_get() failed for kch\nhttpd: Could not reliably determine the server's fully qualified domain name, using 127.0.0.1 for ServerName\n(98)Address already in use: make_sock: could not bind to address [::]:80\n(98)Address already in use: make_sock: could not bind to address 0.0.0.0:80\nno listening sockets available, shutting down\nUnable to open logs\n"}
to retry, use: --limit @/opt/bahmni-installer/bahmni-playbooks/all.retry
Unset `KeepAlive` in the Apache configuration
Experiment
`vi /etc/httpd/conf/httpd.conf`
#
# KeepAlive: Whether or not to allow persistent connections (more than
# one request per connection). Set to "Off" to deactivate.
#
KeepAlive Off
We are facing this issue when doing fresh two machine install (active/passive, master/slave). But, we are not facing this issue in fresh single machine install. You check the pipelines here https://ci.mybahmni.org. Login with GitHub.
One solution that we are thinking of is to replace the httpd restart with httpd reload. That should fix it for now. We are still trying to find out the root cause.
Not sure if I did sudo su, but it does ask for the sudo password upfront - tried several times including the reload, restart change listed above
in any case,
thanks,
Your error looks like because of not running sudo su and when it was trying to install bahmni-emr rpm. The actual httpd error (TASK [httpd-maintenance : reload_httpd_service]) which happens at the end of the installation can be ignored. You can manually restart the service using service httpd restart.
After I tried it again to veryfiy that e6fe6 is included in `0.86-70` it worked out one time. This was when I was doing the update while cooking dinner. So my assumption is that the cause of the problem is that there are too little resources available. I have tried to reproduce this multiple times by removing the external monitor, restarting the system, closing all other applications and don’t touching the system during the update. I had two other successful updates and two failed updates with exacly this scenario.
By the way, here is my system information:
Ubuntu 16.04
16 GB RAM
Intel® Core™ i7-6600U CPU @ 2.60GHz × 4 processor
The VM is set to 2 CPUs, setting it to 4 is discouraged by VirtualBox and also doesn’t make things better if I do it anyway. I have 4096 MB RAM assigned to the VM.
(2) I have logged in with my GitHub accout into go. However, I can only see the build history and not the pipeline definitions. Can you please tell me where to find them and/or grant me read-only access on it?
In case you are interested, 
