Home General

Best Free version of Linux to build a CS/PostGRESQL environment

Hi,
I asked this question several years ago. At the time, the answer was to use Centos Linux. I'm curious today which version others are using to have a Development sandbox that is Linux when you don't have the financial resources to buy the supported OS. In the release notes, Oracle Linux and Red Hat linux are the only "supported" versions (not the free ones as near as I can figure), and only Red Hat Linux for PostGre support (version 10.3). For a development sandbox, I don't need "supported" I just need "won't completely crash when I try to install". What Linux flavours are others using when they need a free Linux variant?
Thanks in advance
-Hugh

Comments

  • Totally not supported, but I have a Debian virtual machine that I use for Linux development. It's just a personal preference of mine. I think most Linux distributions are going to be adequate for development work, as long as you accept that they won't be officially supported.

  • rappardrappard Martin van Rappard EM

    I usually run CentOS since it's closest to RHEL.

    Martin van Rappard
    Technical Support Engineer III
    OpenText

  • I'm guessing that you need to set up both an httpd and a tomcat daemon. I recall the last time I tried this, I discovered that the internal OTDS wasn't there like it is in Windows.
    -Hugh

  • Totally not supported, but I have a Debian virtual machine that I use for Linux development. It's just a personal preference of mine. I think most Linux distributions are going to be adequate for development work, as long as you accept that they won't be officially supported.

    Thanks for sharing Tim Coleman. I'm definitely going to try your preferences. 
  • Back at this now that I have a few days off.
    I'm trying again from scratch with Centos 7.8 and PostGRESQL 10.3 so I can build a CS 20.2 environment. For POSTGRESQL, the requirement are these unaccent and uuid-ossp extensions.  I installed the postgre contributions for PostGRE to get the extensions. I ran
    1. yum install centos-release-scl-rh
    2. Install rh-postgresql10-postgresql-contrib rpm package:
      # yum install rh-postgresql10-postgresql-contrib
    First off, it installed the two control files not in /usr/pg-sql-10/share/extension but in /opt/rh/rh-postgresql10/root/usr/share/pgsql/extension

    So I made a copy of the two .control files into where psql was looking.
    I then get an error like
    ERROR:  could not access file "$libdir/uuid-ossp": No such file or directory
    $libdir in this case resolves to /usr/lib64/pgsql, and in both cases, there is a .so file but it appears to be looking for a *file* named uuid-ossp or unaccent in that directory, and I'm not sure what exactly it's expecting, an empty file or something with content.  It seems to be getting that particular path from the .control file (module_path=$libdir/unaccent for example). The contribution directory had a .sql script. I'm wondering if I should just run the SQL scripts directly rather than trying to create the extension this way.
    Has anyone else run into this in Centos and PSQL 10?
    Thanks in advance
    -Hugh
    PS I googled the **** out of this but couldn't find something that was close enough to my particular use case. Had I more experience with either Linux or PostGRE SQL I might have cracked this by now.

  • To answer my own question, I found that there were two packages that had postgresql-10-contrib. The second package was installed using
    yum install postgresql10-contrib.x86_64
    After that, I was able to create the extensions for both unaccent and uuid-ossp with no issue.
    -Hugh

  • I'm further along, although I have to admin it really doesn't seem like there has been much QA on the Linux installation.  I get as far as establishing the connection (I use a pre-created db and user that has ownership on the db). The credentials are accepted, and I specify a doc store, but after a minute or so, I get an SDNR error message but no stack trace and nothing untoward showing in the log file (except it looks like the server restarted itself). I was previously getting a 504 Gateway timeout which I thought was due to the long time it takes to create the db.  I fixed that by setting a 5 minute timeout in the Apache httpd.conf file.  I've verified the db tables did get created although it didn't populate any of the system objects yet.  Do others working with Linux end up hitting this as well?  How do you get to the next step?  I still need to create a Content Server resource in my OTDS server but I need to get this moved further along.
    -Hugh 
  • I managed to complete the build today YAY!!!! Two tricks I used to get to make this work are as follows:
    1) in the httpd.conf file, I set TimeOut to 300 (5 minutes);
    2) I started the config with the admin server stopped so that I would get a message indicating that I need to start it (however I think that the tables do get built there);
    I also tailed the thread1.out file to monitor progress. On the request handler admin.checkAdminSvr which required the start_lladmin (*livelink* admin :) ), the thread was stuck there for about 6 minutes. When the thread ended, I got another error message saying that I got an empty response.  Not perturbed, at that point, I just entered the URL for what I saw as the nextURL pattern which was the admin.setup request handler. That got me back on track.
    One small additional hiccup I hit was on the licence file.  If you downloaded it in Windows, Linux won't be able to read it, so download again in Linux.
    I'm curious if more clients are looking at using Linux and PostGRESQL. I know that one past client was eager to switch because the Oracle licensing was killing them.  I certainly like the hands-on nature of Linux so I'm hoping more clients go this route. Windows really dumbs things down :/
    -Hugh
  • @Tim_Coleman does your Ubuntu environment have an OTDS server on it? I just tried the installation on a new Ubuntu 20.4 with OpenJDK 11, and Tomcat 9.0.37 and the installer crapped out with error code 1. Never had this on Centos, so I'll stick with Centos 7 for Linux Content Server for now, and consider Ubuntu for a replacement desktop when the time comes (I don't want to go to Windows 10 due to its invasive nature of logging data without any choice in the matter).
    -Hugh
  • I'm further along, although I have to admin it really doesn't seem like there has been much QA on the Linux installation.  I get as far as establishing the connection (I use a pre-created db and user that has ownership on the db). The credentials are accepted, and I specify a doc store, but after a minute or so, I get an SDNR error aol mail login message but no stack trace and nothing untoward showing in the log file (except it looks like the server restarted itself). I was previously getting a 504 Gateway timeout which I thought was due to the long time it takes to create the db.  I fixed that by setting a 5 minute timeout in the Apache httpd.conf file.  I've verified the db tables did get created although it didn't populate any of the system objects yet.  Do others working with Linux end up hitting this as well?  How do you get to the next step?  I still need to create a Content Server resource in my OTDS server but I need to get this moved further along. 
    -Hugh 
    Totally agreed.

  • Revisiting this a couple of months later.  I fired up my Centos CS 20.2 instance and I'm noticing that on restarting, the llserver process never starts. No error message, no nothing.  Has anyone working with Centos found this?  I thought it might be firewall related so I added a rule for port 2099.
      In a related note, if anyone doing a Centos 7 sandbox is having issues with Apache Tomcat, make sure that the openDJ service is running. Unlike with the Windows build, OpenDJ doesn't start automagically.
    -Hugh
  • Glad you like it. We have been using this for several weeks now and it seems to be going nicely MyLowesLife.
Sign In or Register to comment.