Slurm package conflict with yum update
Slurm Resource Manager for Rocks Clusters
Brought to you by:
wernsaar
Menu
▾
▴
Hi Werner,
When I run
yum update, I get the following error.... --> Processing Dependency: libprotobuf.so.8()(64bit) for package: mosh-1.3.0-1.el7.x86_64 ---> Package perl-Socket6.x86_64 0:0.23-15.el7 will be installed ---> Package volume_key-libs.x86_64 0:0.3.9-8.el7 will be installed --> Finished Dependency Resolution Error: Package: slurm-17.11.5-1.el7.centos.x86_64 (@Rocks-7.0) Requires: libipmimonitoring.so.5()(64bit) Removing: freeipmi-1.4.11-1.el7.centos.x86_64 (@Rocks-7.0) libipmimonitoring.so.5()(64bit) Updated By: freeipmi-1.5.7-2.el7.x86_64 (base) ~libipmimonitoring.so.6()(64bit) Error: Package: mosh-1.3.0-1.el7.x86_64 (epel) Requires: libprotobuf.so.8()(64bit) Available: protobuf-2.5.0-8.el7.x86_64 (base) libprotobuf.so.8()(64bit) Installed: protobuf-2.6.1-2.x86_64 (@anaconda/7.0) ~libprotobuf.so.9()(64bit) You could try using --skip-broken to work around the problem You could try running: rpm -Va --nofiles --nodigestDo you have any idea?
Hi,
the ISO
https://sourceforge.net/projects/slurm-roll/files/release-7.0.0-17.11.05/slurm-7.0.0-17.11.05.00.02.x86_64.disk1.iso
solves the problem with old FreeIPMI packages.
But your problem is, that you have enabled the Epel repository (I never
do this),
but mosh is already installed from Centos Base.
Best regards
Werner
On 11/07/2018 02:19 PM, mahmood naderan wrote:
When I disable epel, I get
---> Package freeipmi.x86_64 0:1.4.11-1.el7.centos will be updated --> Processing Dependency: libipmimonitoring.so.5()(64bit) for package: slurm-17.11.5-1.el7.centos.x86_64 ---> Package libbytesize.x86_64 0:1.2-1.el7 will be installed ---> Package volume_key-libs.x86_64 0:0.3.9-8.el7 will be installed --> Finished Dependency Resolution Error: Package: slurm-17.11.5-1.el7.centos.x86_64 (@Rocks-7.0) Requires: libipmimonitoring.so.5()(64bit) Removing: freeipmi-1.4.11-1.el7.centos.x86_64 (@Rocks-7.0) libipmimonitoring.so.5()(64bit) Updated By: freeipmi-1.5.7-2.el7.x86_64 (base) ~libipmimonitoring.so.6()(64bit) You could try using --skip-broken to work around the problem You could try running: rpm -Va --nofiles --nodigestIt seems that the freeimpi package comes from rocks repository
Attempt to remove the package will result in uninstalling slurm which is not what I want
It seems that this is a problem with rocks repo. Do you agree?
Also, trying to update the package, it ends up with error regarding the slurm package.
--> Finished Dependency Resolution Error: Package: slurm-17.11.5-1.el7.centos.x86_64 (@Rocks-7.0) Requires: libipmimonitoring.so.5()(64bit) Removing: freeipmi-1.4.11-1.el7.centos.x86_64 (@Rocks-7.0) libipmimonitoring.so.5()(64bit) Updated By: freeipmi-1.5.7-2.el7.x86_64 (base) ~libipmimonitoring.so.6()(64bit)Should I install the latest version of slurm roll first?
The latest version is 18.08 which is a major upgrade so you might prefer to upgrade to the last release in the 17.11 series: 17.11.09.02.01
To avoid conflicts with epel you can exclude packages in /etc/yum.repos.d/epel.repo:
Right, but as I said, by excluding mosh, still libipmi has some conflicts. I don't know is it safe to remove current slurm and install the latest one, or update the current to the latest version because I see the following statement in the manual
Mine is now slurm-7.0.0-17.11.03.02.02.x86_64.disk1.iso
Last edit: mahmood naderan 2018-11-12
Hi,
you can safely update to the slurm versions 17.11.9-2 or 18.08.3
On 11/12/2018 08:21 AM, mahmood naderan wrote:
So, the process of updating the frontend was successfull.
[root@rocks7 ~]# rocks add attr slurm_sync_enable false [root@rocks7 ~]# export LANG=C [root@rocks7 ~]# rocks disable roll slurm [root@rocks7 ~]# rocks add roll slurm-7.0.0.185-18.08.03.01.00.x86_64.disk1.iso mount: /dev/loop0 is write-protected, mounting read-only Copying slurm to Rolls.....45659 blocks [root@rocks7 ~]# rocks enable roll slurm [root@rocks7 ~]# cd /export/rocks/install/ [root@rocks7 install]# rocks create distro Cleaning distribution Resolving versions (base files) including "core" (7.0,x86_64) roll... including "CentOS" (7.4.1708,x86_64) roll... including "kernel" (7.0,x86_64) roll... including "python" (7.0,x86_64) roll... including "slurm" (7.0.0.185,x86_64) roll... including "Updates-CentOS-7.4.1708" (2017-12-01,x86_64) roll... including "perl" (7.0,x86_64) roll... including "hpc" (7.0,x86_64) roll... including "base" (7.0,x86_64) roll... including "ganglia" (7.0,x86_64) roll... Including critical RPMS Resolving versions (RPMs) including "core" (7.0,x86_64) roll... including "CentOS" (7.4.1708,x86_64) roll... including "kernel" (7.0,x86_64) roll... including "python" (7.0,x86_64) roll... including "slurm" (7.0.0.185,x86_64) roll... including "Updates-CentOS-7.4.1708" (2017-12-01,x86_64) roll... including "perl" (7.0,x86_64) roll... including "hpc" (7.0,x86_64) roll... including "base" (7.0,x86_64) roll... including "ganglia" (7.0,x86_64) roll... Resolving versions (SRPMs) including "core" (7.0,x86_64) roll... including "CentOS" (7.4.1708,x86_64) roll... including "kernel" (7.0,x86_64) roll... including "python" (7.0,x86_64) roll... including "slurm" (7.0.0.185,x86_64) roll... including "Updates-CentOS-7.4.1708" (2017-12-01,x86_64) roll... including "perl" (7.0,x86_64) roll... including "hpc" (7.0,x86_64) roll... including "base" (7.0,x86_64) roll... including "ganglia" (7.0,x86_64) roll... Creating files (symbolic links - fast) Applying netstage (aka stage2) Applying rocks-anaconda-updates Installing XML Kickstart profiles installing "ganglia" profiles... installing "perl" profiles... installing "base" profiles... installing "python" profiles... installing "hpc" profiles... installing "slurm" profiles... installing "kernel" profiles... installing "core" profiles... installing "site" profiles... Calling Yum genpkgmetadata.py Creating repository Rebuilding Product Image including md5 sums Creating Directory Listing [root@rocks7 install]# yum clean all Loaded plugins: fastestmirror, langpacks Cleaning repos: Rocks-7.0 base epel Cleaning up everything Maybe you want: rm -rf /var/cache/yum, to also free up space taken by orphaned data from disabled or removed repos Cleaning up list of fastest mirrors [root@rocks7 install]# yum update ... ... Complete! You have new mail in /var/spool/mail/root [root@rocks7 install]# sinfo -version sinfo: error: Error: invalid entry for --iterate=on [root@rocks7 install]# sinfo --version slurm 18.08.3Heading for the computes.
Hi, I'm new, I'm doing a Rocks 7.0 installation like this in the rocks installation manual, up to here everything is correct.
I have a node installed and I work correctly.
The roll I have installed are:
base: 7.0 x86_64 yes
CentOS: 7.4.1708 x86_64 yes
core: 7.0 x86_64 yes
kernel: 7.0 x86_64 yes
Updates-CentOS-7.4.1708: 2017-12-01 x86_64 yes
my doubt is that I need more to start installing Slurm?
The steps that followed are those found in the manual for installation:
export LANG = C
rocks add roll slurm * .iso
rocks enable roll slurm
cd / export / rocks / install
rocks create distro
yum clean all (* here I have a problem: Maybe you want: rm -rf / var / cache / yum, to also free up space taken by orphaned data from disabled or removed repos
Cleaning up list of fastest mirrors
) * I already tried to remove and I did not have results.
yum update
rocks run roll slurm | sh
(Here is this:
Package 1: mariadb-5.5.56-2.el7.x86_64 already installed and latest version
Package 1: mariadb-5.5.56-2.el7.x86_64 already installed and latest version
Package pdsh-2.31-1.el7.x86_64 already installed and latest version
No package slurm-munge available.
Resolving Dependencies
as I correct it * I want to think that in some repository. )
then all these errors come out. some help please.
/ bin / mkdir: can not create directory '/ etc / slurm': File exists
Created symlink from /etc/systemd/system/multi-user.target.wants/munge.service to /usr/lib/systemd/system/munge.service.
Created symlink from /etc/systemd/system/multi-user.target.wants/slurmd.service to /usr/lib/systemd/system/slurmd.service.
/etc/logrotate.d/RCS/slurm,v <- /etc/logrotate.d/slurm
Initial revision: 1.1
done
/etc/logrotate.d/RCS/slurm,v -> /etc/logrotate.d/slurm
revision 1.1 (locked)
done
/ bin / mkdir: can not create directory '/ etc / slurm': File exists
/etc/slurm/RCS/suspendhost.sh,v <- /etc/slurm/suspendhost.sh
Initial revision: 1.1
done
/etc/slurm/RCS/suspendhost.sh,v -> /etc/slurm/suspendhost.sh
revision 1.1 (locked)
done
/etc/slurm/RCS/resumehost.sh,v <- /etc/slurm/resumehost.sh
Initial revision: 1.1
done
/etc/slurm/RCS/resumehost.sh,v -> /etc/slurm/resumehost.sh
revision 1.1 (locked)
done
Created symlink from /etc/systemd/system/multi-user.target.wants/slurmdbd.service to /usr/lib/systemd/system/slurmdbd.service.
Created symlink from /etc/systemd/system/multi-user.target.wants/mariadb.service to /usr/lib/systemd/system/mariadb.service.
sacctmgr: error: slurm_persist_conn_open_without_init: failed to open persistent connection to 172: 6819: Invalid argument
sacctmgr: error: slurmdbd: Sending PersistInit msg: Invalid argument
sacctmgr: error: Problem talking to the database: Invalid argument
sacctmgr: error: slurm_persist_conn_open_without_init: failed to open persistent connection to 172: 6819: Invalid argument
sacctmgr: error: slurmdbd: Sending PersistInit msg: Invalid argument
sacctmgr: error: Problem talking to the database: Invalid argument
sacctmgr: error: slurm_persist_conn_open_without_init: failed to open persistent connection to 172: 6819: Invalid argument
sacctmgr: error: slurmdbd: Sending PersistInit msg: Invalid argument
sacctmgr: error: Problem talking to the database: Invalid argument
sacctmgr: error: slurm_persist_conn_open_without_init: failed to open persistent connection to 172: 6819: Invalid argument
sacctmgr: error: slurmdbd: Sending PersistInit msg: Invalid argument
sacctmgr: error: Problem talking to the database: Invalid argument
sacctmgr: error: slurm_persist_conn_open_without_init: failed to open persistent connection to 172: 6819: Invalid argument
sacctmgr: error: slurmdbd: Sending PersistInit msg: Invalid argument
sacctmgr: error: Problem talking to the database: Invalid argument
sacctmgr: error: slurm_persist_conn_open_without_init: failed to open persistent connection to 172: 6819: Invalid argument
sacctmgr: error: slurmdbd: Sending PersistInit msg: Invalid argument
sacctmgr: error: Problem talking to the database: Invalid argument
Created symlink from /etc/systemd/system/multi-user.target.wants/slurmctld.service to /usr/lib/systemd/system/slurmctld.service.
Job for slurmctld.service failed because a configured resource was exceeded. See "systemctl status slurmctld.service" and "journalctl -xe" for details.
##############################################################################################
WARNING: The command:
sacctmgr -i create cluster test
failed Please run this command again
##############################################################################################
reboot
Could you tell me what repositories do I need to enable or include?
The Slurm version is release-7.0.0-18.08.03.
Please Download and read slurm-roll.pdf.
Install all required rolls and don't install the roll area51.
And Check the downloaded iso with md5sum. It may be corrupt.
On 12/01/2018 12:13 AM, armando martinez bautista wrote:
do not enable external repositories like epel
On 12/01/2018 05:44 AM, Werner Saar wrote:
Yes thanks
base:
ganglia
hpc:
nucleus:
nucleus:
CentOS:
Updates-CentOS-7.4.1708:
I'm doing a new installation of Rocks 7.0 and I just install those roll.
in the host name (localhost.localdomain) I have replaced it with the ip of the cluster.
I did the slurm installation again.
I have the following problem
"No package slurm-munge available."
Now, can you send me the repository link?
I have the address as baseurl = http: // (ip-del-cluster) / install / rocks-dist / x86_64
Maybe this more from there.
If anyone has the correct repository link, would I thank them, or do I try to install the slurm-munge separately?
do not activate * epel.