# Installation on Centos 8 Stream **URL:** \ **Category:** Users/Site Administrators\ **Created:** [July 6, 2021, 12:18pm UTC](https://community.openpbs.org/t/installation-on-centos-8-stream/2689 "2021-07-06T12:18:36Z")\ **Posts on this page:** 9\ **Page:** 1
**Author:** ![boboshaq](https://avatars.discourse-cdn.com/v4/letter/b/bc8723/32.png) [@boboshaq](https://community.openpbs.org/u/boboshaq)\ **Post date:** [July 6, 2021, 12:18pm UTC](https://community.openpbs.org/t/installation-on-centos-8-stream/2689/1 "2021-07-06T12:18:36Z")
**Is anyone was installed properly Openpbs 20 on Centos8 Stream ???** ---
**Author:** ![speleolinux](https://avatars.discourse-cdn.com/v4/letter/s/779978/32.png) [@speleolinux](https://community.openpbs.org/u/speleolinux)\ **Post date:** [July 7, 2021, 2:47am UTC](https://community.openpbs.org/t/installation-on-centos-8-stream/2689/2 "2021-07-07T02:47:29Z")
$ cat /etc/centos-release CentOS Linux release 8.4.2105 Not sure this info will help you as I have not used OpenPBS but PBSPro 20.1.2 does work. I used the rpms to install. I would expect that it should work fine. Mike ---
**Author:** ![boboshaq](https://avatars.discourse-cdn.com/v4/letter/b/bc8723/32.png) [@boboshaq](https://community.openpbs.org/u/boboshaq)\ **Post date:** [July 7, 2021, 7:30am UTC](https://community.openpbs.org/t/installation-on-centos-8-stream/2689/3 "2021-07-07T07:30:25Z")
Hi, Thanks for info. It’s seems to openpbs and pbs pro working on Centos 8 but not working on CentOS Stream release 8 ☹ ---
**Author:** ![alexis.cousein](https://yyz2.discourse-cdn.com/flex030/user_avatar/community.openpbs.org/alexis.cousein/32/265_2.png) [@alexis.cousein](https://community.openpbs.org/u/alexis.cousein)\ **Post date:** [July 8, 2021, 9:01am UTC](https://community.openpbs.org/t/installation-on-centos-8-stream/2689/4 "2021-07-08T09:01:28Z")
Perhaps it would be more illuminating to post exactly what you’re doing and what the error is? The simple answer to your question is: yes, people have compiled and used OpenPBS on CentOS8 stream, but I doubt that’s what you’re after. It’s possible precompiled RPMs don’t work; that’s usually fixable too, but again: what is the problem? ---
**Author:** ![boboshaq](https://avatars.discourse-cdn.com/v4/letter/b/bc8723/32.png) [@boboshaq](https://community.openpbs.org/u/boboshaq)\ **Post date:** [July 8, 2021, 9:25am UTC](https://community.openpbs.org/t/installation-on-centos-8-stream/2689/5 "2021-07-08T09:25:33Z")
Hi @alexis.cousein , The first problem was with installation openpbs via rpm packages. Problem with missing dependency libhwloc.so.5. Problem was described here: > [@Problem with dependency hwloc openpbs20.0.1](https://community.openpbs.org/t/problem-with-dependency-hwloc-openpbs20-0-1/2656): > > Hi, I trying install package: openpbs-execution-20.0.1-0.x86\_64.rpm on Centos8-stream and I have one error: error: Failed dependencies: libhwloc.so.5()(64bit) is needed by openpbs-execution-20.0.1-0.x86\_64 I have installed hwloc packages: Installed Packages hwloc.x86\_64 2.2.0-2.el8 hwloc-devel.x86\_64 2.2.0-2.el8 hwloc-libs.x86\_64 2.2.0-2.el8 hwloc-plugins.x86\_64 2.2.0-2.el8 They are pretty old according this webpage: [https:/…](https://www.open-mpi.org/software/hwloc/ancient/) I solved it compiling openpbs from sources. Second, unresolved problem was with communication between server and nodes. It is described here: > [@State-unknown,down](https://community.openpbs.org/t/state-unknown-down-on-centos-8-stream/2681): > > Is anyone installed properly openpbs 20 on Centos 8 Stream? Nodes are still in unknown,down state. Regards! ---
**Author:** ![saxton](https://avatars.discourse-cdn.com/v4/letter/s/ecccb3/32.png) [@saxton](https://community.openpbs.org/u/saxton)\ **Post date:** [August 3, 2021, 2:48am UTC](https://community.openpbs.org/t/installation-on-centos-8-stream/2689/6 "2021-08-03T02:48:49Z")
I encounter several problems installing openpbs-20.0.1 on “CentOS Stream 8” with AMD architecture. Following the options for “CentOS-8” in the “INSTALL” instructions, my first problem begins at step 2 during the installation of the prerequisites “postgresql-\*”. I am repeatedly warned about a non-existent user and group: > warning: user postgres does not exist - using root > warning: group postgres does not exist - using root and later I am warned: > [/usr/lib/tmpfiles.d/postgresql.conf:1] Unknown user ‘postgres’. > warning: %transfiletriggerin(systemd-239-49.el8.x86\_64) scriptlet failed, exit status 65 > > Error in scriptlet in rpm package postgresql-server > Verifying : postgresql-server-10.17-2.module\_el8.5.0+865+7313c56 1/1 Installed products updated. > > Installed: > postgresql-server-10.17-2.module\_el8.5.0+865+7313c562.x86\_64 If I ignore these warnings, then the PBS services fail to start at Step 13. I could manually create a user “postgres” and provide a group, but that would be futile if something important is missing from “postgresql-server” or from “openpbs-20.0.1”. Do “postgresql-server” and “postgresql-contrib” need something extra? Could someone please suggest how to improve the installation process on “CentOS Stream 8”? ---
**Author:** ![saxton](https://avatars.discourse-cdn.com/v4/letter/s/ecccb3/32.png) [@saxton](https://community.openpbs.org/u/saxton)\ **Post date:** [August 3, 2021, 2:56am UTC](https://community.openpbs.org/t/installation-on-centos-8-stream/2689/7 "2021-08-03T02:56:17Z")
My next installation problem seems similar. Later on, I receive warnings and errors because of the non-existence of a user “smmsp” in group “smmsp”. I guess that this user is part of “sendmail.” Shouldn’t these users be created automatically when they’re needed during the OpenPBS installation recipe? Does my operating system have some wrong settings that prevent scripted user creation? ---
**Author:** ![saxton](https://avatars.discourse-cdn.com/v4/letter/s/ecccb3/32.png) [@saxton](https://community.openpbs.org/u/saxton)\ **Post date:** [August 3, 2021, 11:16pm UTC](https://community.openpbs.org/t/installation-on-centos-8-stream/2689/8 "2021-08-03T23:16:38Z")
I confirm that it is possible to install and run OpenPBS under CentOS-Stream-8, at least under some circumstances (AMD Threadripper). My particular installation problems are solved. Let me summarise, for the benefit of anyone who faces the same issues. Installing the packages “postgresql” and “sendmail” (in steps 1 and 2 of official “INSTALL” instructions) should have created users “postgres” and “smmsp”, but failed to do so (on my system). It is sufficient to create these users manually, after the “yum” and “dnf” commands, but before compiling or executing the PBS post-install. For the sake of cleanliness, I delete any old versions of these users that might be left over from previous installation attempts. > sudo userdel -r postgres > sudo userdel -r smmsp > sudo rm -rf /home/postgres /home/smmsp > > sudo dnf install postgresql-devel postgresql-contrib > … > sudo dnf install sendmail > > sudo useradd -m postgres > sudo useradd -m smmsp After fixing the critical error of missing users, the “postgresql” packages can be installed with just a few further warnings. I don’t know whether these warnings would occur generally or are specific to my machine and its type. The warnings were: > [/usr/lib/tmpfiles.d/pesign.conf:1] Line references path below legacy directory /var/run/, updating /var/run/pesign → /run/pesign; please update the tmpfiles.d/ drop-in file accordingly. > [/usr/lib/tmpfiles.d/postgresql.conf:1] Line references path below legacy directory /var/run/, updating /var/run/postgresql → /run/postgresql; please update the tmpfiles.d/ drop-in file accordingly. I edited these two files to remove the offensive substring “/var”, from the first line of each. > sudo vi /usr/lib/tmpfiles.d/postgresql.conf > sudo vi /usr/lib/tmpfiles.d/pesign.conf After performing these interventions, I now have OpenPBS 20.0.1 installed and running smoothly (CentOS-Stream-8; AMD Threadripper cpu). ---
**Author:** ![saxton](https://avatars.discourse-cdn.com/v4/letter/s/ecccb3/32.png) [@saxton](https://community.openpbs.org/u/saxton)\ **Post date:** [February 14, 2022, 6:55pm UTC](https://community.openpbs.org/t/installation-on-centos-8-stream/2689/9 "2022-02-14T18:55:41Z")
I was running OpenPBS 20.0.1 on a “CentOS Stream 8” computer successfully for half a year, until a recent update of the operating system. Now I face multiple problems when I attempt to reinstall and start OpenPBS. The current system is “Linux 4.18.0-365.el8.x86\_64” with “AMD threadripper pro 3995wx” architecture. The changes were great enough that my “Timeshift” application cannot revert to the previous system state. Postgresql version 12 is reinstalled and running OK. Here, I would like to ask about the earliest problems that I encounter when reinstalling OpenPBS 20.0.1 from the source. I am following these official instructions: > [https://raw.githubusercontent.com/openpbs/openpbs/master/INSTALL](https://raw.githubusercontent.com/openpbs/openpbs/master/INSTALL) **Problem 1.** At step 4 of the instructions, I run `./autogen.sh` and receive many errors like this: > `test/tests/Makefile.am:60: while processing 'dist_ptl_testselftest_DATA'` > `automake: warning: wildcard $(srcdir: non-POSIX variable name` > `automake: (probably a GNU make extension)` I add one line to the end of `configure.ac` > `AC_SUBST([DOLLAR],[$])` and replace every instance of > `$(wildcard` with > `@DOLLAR@(wildcard` in the files `test/tests/Makefile.am` and `test/fw/Makefile.am`. Are there any other Makefiles that need changing? This remedy was inspired by this link: > [https://stackoverflow.com/questions/15711824/autoreconf-stops-with-non-posix-variable-name](https://stackoverflow.com/questions/15711824/autoreconf-stops-with-non-posix-variable-name) **Problem 2.** Running `./autogen.sh` again reveals a new type of error message. The last instance says: > `src/tools/Makefile.am:219: warning: source file '$(top_srcdir)/src/lib/Libcmds/cmds_common.c' is in a subdirectory, src/tools/Makefile.am:219: but option 'subdir-objects' is disabled` I change one line in `configure.ac` from > ` [*], m4_define([am_init_string], [-Wall foreign]))` to > ` [*], m4_define([am_init_string], [-Wall foreign subdir-objects]))` after taking inspiration from: > [https://stackoverflow.com/questions/38722465/automake-subdir-objects-is-disabled](https://stackoverflow.com/questions/38722465/automake-subdir-objects-is-disabled) After these changes, `autogen.sh` doesn’t give me any further warnings or errors. These are my messages resulting from step 4 in the official instructions: > `Generating configure script and Makefile templates.` > `libtoolize: putting auxiliary files in AC_CONFIG_AUX_DIR, 'buildutils'.` > `libtoolize: copying file 'buildutils/ltmain.sh'` > `libtoolize: putting macros in AC_CONFIG_MACRO_DIRS, 'm4'.` > `libtoolize: copying file 'm4/libtool.m4'` > `libtoolize: copying file 'm4/ltoptions.m4'` > `libtoolize: copying file 'm4/ltsugar.m4'` > `libtoolize: copying file 'm4/ltversion.m4'` > `libtoolize: copying file 'm4/lt~obsolete.m4'` > `configure.ac:61: installing 'buildutils/compile'` > `configure.ac:60: installing 'buildutils/missing'` > `src/cmds/Makefile.am: installing 'buildutils/depcomp'` I suffer other problems later on, when I try to start the compiled pbs service. For now, I would appreciate advice or confirmation that my changes to help `autogen.sh` are correct, necessary, and sufficient. Please check the “sanity” of my tinkering. Is there a simple reason why `autogen.sh` behaved so badly on a straightforward CentOS machine? Is there a more elegant remedy possible?