merge bug fixes found in beta into master branch (#96)
fix:the size of interactive job working place cannot be resized #74 (#75)
fix: prevent mbatchd core dump when bsub -pack jobs use -f (#77)
fix: incorrect resreq display and unit conversion when LSF_UNIT_FOR_LIMITS is set (#79)
- Fix incorrect effective/combined resreq display when LSF_UNIT_FOR_LIMITS changes the unit
- Fix unit conversion to MB: use 1024 instead of 1000
- Refactor LSF_UNIT_FOR_LIMITS-related code to simplify the logic
docs: document the bmod esub option (#88)
docs: fix README user guide link (#84)
fix: resolve installer helper from script directory (#82)
upgrade to 2.2.0 release from 2.2.0 beta
Co-authored-by: tczz88888 939948564@qq.com Co-authored-by: Wu Shuwen mikewushuwen@outlook.com
版权所有:中国计算机学会技术支持:开源发展技术委员会
京ICP备13000930号-9
京公网安备 11010802047560号
Description
volclava, whose latest version is 2.2.0, is developed based on OpenLava 2.0 with following its license of GPL 2.0.
OpenLava is a free and open-source workload scheduler that’s fully compatible with IBM LSF. It supports diverse high-performance computing and analytical applications. It came from an early version of LSF through open-sourcing. Its command-line and file formats are similar to most LSF functions, so both users and administrators can easily get the hang of operating it.
volclava fixes obvious flaws and adds necessary features, covering the basic functional needs of the EDA business. It’s recommended for use in small-scale clusters (with less than 100 nodes) where scheduling performance requirements aren’t too high.
Everyone is welcomed to feed back via git issue.
Latest Version
Volclava 2.2.0 - Public Release
This version is built from the master branch with source code tagged as
2.2.0.For detailed release notes and updates, please visit: Release Notes
Support OS
The development and testing operating system for the volclava project 2.2 is CentOS Linux release 7.9.2009 (Core), which is also one of the commonly used operating system versions in IC design. CentOS 6, CentOS 7, CentOS 8, as well as their corresponding Red Hat and Rocky versions should all be able to run. The main potential risk lies in that differences in system library versions may affect the operation of some components.
Install
The following uses the setup of a 3-node cluster (master-test,cmp1-test,cmp2-test) as an example. Please confirm that all hosts in the volclava cluster can access the top-level installation directory of volclava using the same path name. For the convenience of management, it is recommended to move the software to the shared storage path
Download the code onto the master node
Log on master node as root, and enter the source code package directory
Create an install.conf file to customize the cluster setup
Edit the file, then save it with the following content:
The default administrator account and cluster name are volclava. The following demonstrations will use these values, but you can specify your own.
Run install script ``` [root@master-test volclava]# ./volcinstall.sh –file=./install.conf …. make[1]: Leaving directory `/install_pkg/volclava’ make: warning: Clock skew detected. Your build may be incomplete. Congratulates, installation is done and enjoy the journey!
[root@cmp1-test ~]# /install_pkg/volclava/volcinstall.sh –type=server –env=/software/volclava-2.2 [root@cmp2-test ~]# /install_pkg/volclava/volcinstall.sh –type=server –env=/software/volclava-2.2
[root@master-test ~]# cat /etc/hosts 127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4 ::1 localhost localhost.localdomain localhost6 localhost6.localdomain6 192.168.1.1 master-test 192.168.1.2 cmp1-test 192.168.1.3 cmp2-test
[root@master-test-test etc]# vim /software/volclava-2.2/etc/lsf.cluster.volclava
[root@master-test ~]# yum install psmisc [root@master-test ~]# service volclava restart [root@cmp1-test ~]# yum install psmisc [root@cmp1-test ~]# service volclava restart [root@cmp2-test ~]# yum install psmisc [root@cmp2-test ~]# service volclava restart
[root@master-test ~]# service volclava status ##check daemons status lim pid: <5922> res pid: <5924> sbatchd pid: <5927> lim mbatchd: <5940>
[root@master-test ~]# lsid ##check cluster status volclava project 2.2, Nov 11 2025
My cluster name is volclava My master name is master-test
[root@cmp2-test etc]# lshosts ##check cluster static resources HOST_NAME type model cpuf ncpus maxmem maxswp server RESOURCES master-test linux IntelI5 100.0 4 32012M 10239M Yes (cs) cmp1-test linux IntelI5 100.0 4 31993M 10239M Yes (cs) cmp2-test linux IntelI5 100.0 4 31993M 10239M Yes (cs)
[root@cmp2-test etc]# lsload ##check cluster load HOST_NAME status r15s r1m r15m ut pg ls it tmp swp mem master-test ok 0.0 0.0 0.1 0% 0.0 1 0 3321M 10G 28G cmp1-test ok 0.0 0.0 0.0 0% 0.0 1 6 20G 10G 29G cmp2-test ok 0.0 0.0 0.0 0% 0.0 1 24 45G 10G 30G
[root@cmp2-test etc]# bhosts ##check workload on hosts HOST_NAME STATUS JL/U MAX NJOBS RUN SSUSP USUSP RSV cmp1-test ok - 4 0 0 0 0 0 cmp2-test ok - 4 0 0 0 0 0 master-test ok - 4 0 0 0 0 0
[root@master-test ~]# su - volclava [volclava@master-test ~]bsubsleep100Job<1>issubmittedtodefaultqueue<normal>.[volclava@master−test ] bjobs JOBID USER STAT QUEUE FROM_HOST EXEC_HOST JOB_NAME SUBMIT_TIME 1 volclav PEND normal master-test sleep 100 Nov 27 15:03