我一直在使用this guide在我的节点上安装Slurm。我仅尝试将slurmd示例复制到slurm.conf并按照指南运行sudo start slurmd
,这会在journalctl_ per中生成以下错误输出:
Dec 29 19:16:22 Node_2 slurmd[27681]: error: plugin_load_from_file: dlopen(/usr/lib/slurm/select_linear.so): /usr/lib/slurm/select_linear.so: undefined symbol: slurm_job_preempt_mode
Dec 29 19:16:22 Node_2 slurmd[27681]: error: Couldn't load specified plugin name for select/linear: Dlopen of plugin file failed
Dec 29 19:16:22 Node_2 systemd[1]: slurmd.service: Control process exited, code=exited status=1
Dec 29 19:16:22 Node_2 slurmd[27681]: fatal: Can't find plugin for select/linear
Dec 29 19:16:22 Node_2 systemd[1]: slurmd.service: Failed with result 'exit-code'.
Dec 29 19:16:22 Node_2 systemd[1]: Failed to start Slurm node daemon.
但是,我确实有这个插件,我可以看到它:
sudo ls /usr/lib/slurm/select_linear.so
/usr/lib/slurm/select_linear.so
我还尝试了sudo slurmd -cDvvvvv
,并得到了以下输出。
slurmd: error: plugin_load_from_file: dlopen(/usr/lib/slurm/select_linear.so): /usr/lib/slurm/select_linear.so: undefined symbol: slurm_job_preempt_mode
slurmd: error: Couldn't load specified plugin name for select/linear: Dlopen of plugin file failed
slurmd: fatal: Can't find plugin for select/linear
我还尝试了sudo slurmctld -cDvvvvv
,得到的结果如下:
slurmctld: debug: Log file re-opened
slurmctld: debug: creating clustername file: /var/spool/slurm/ctld/clustername
slurmctld: Stack size set to 8388608
slurmctld: slurmctld version 17.11.0 started on cluster linux
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/crypto_munge.so
slurmctld: Munge cryptographic signature plugin loaded
slurmctld: debug3: Success.
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/select_linear.so
slurmctld: debug3: Success.
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/preempt_none.so
slurmctld: preempt/none loaded
slurmctld: debug3: Success.
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/checkpoint_none.so
slurmctld: debug3: Success.
slurmctld: debug: Checkpoint plugin loaded: checkpoint/none
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/acct_gather_energy_none.so
slurmctld: debug: AcctGatherEnergy NONE plugin loaded
slurmctld: debug3: Success.
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/acct_gather_profile_none.so
slurmctld: debug: AcctGatherProfile NONE plugin loaded
slurmctld: debug3: Success.
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/acct_gather_interconnect_none.so
slurmctld: debug: AcctGatherInterconnect NONE plugin loaded
slurmctld: debug3: Success.
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/acct_gather_filesystem_none.so
slurmctld: debug: AcctGatherFilesystem NONE plugin loaded
slurmctld: debug3: Success.
slurmctld: debug2: No acct_gather.conf file (/etc/slurm-llnl/acct_gather.conf)
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/jobacct_gather_none.so
slurmctld: debug: Job accounting gather NOT_INVOKED plugin loaded
slurmctld: debug3: Success.
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/ext_sensors_none.so
slurmctld: ExtSensors NONE plugin loaded
slurmctld: debug3: Success.
slurmctld: debug3: Trying to load plugin /usr/lib/slurm/switch_none.so
slurmctld: debug: switch NONE plugin loaded
slurmctld: debug3: Success.
slurmctld: debug: power_save module disabled, SuspendTime < 0
slurmctld: error: this host (Node_2/Node_2) not a valid controller (linux0 or (null))
你知道我需要做什么才能看到这个插件吗?
发布于 2018-01-07 17:42:41
在我看来,这看起来像是SLURM中的一个bug。
我的猜测是,select/linear
插件只有在被slurmctld
使用时才有意义,而在被slurmd
使用时没有意义。slurm_job_preempt_mode
符号确实是在slurmctld
中定义的,但在slurmd
中没有定义。
FWIW是一个稍微老一点的版本,在slurmd
中有相同的“缺失”符号,在RHEL7上运行得很好,所以我猜行为取决于操作系统的链接器(配置)。
我最好的选择是你直接向SLURM的人报告这个问题。
发布于 2018-04-27 15:01:03
我用另一个插件(lua)也遇到了同样的问题。我的解决方案是在rpmbuild上提到它:
rpmbuild --with lua -ta slurm-16.05.7.tar.bz2
也许这也会对你有帮助,我会试着这样做
rpmbuild --with linear -ta slurm-16.05.7.tar.bz2
https://stackoverflow.com/questions/48070840
复制相似问题