This document proposes a broad change to NRAO’s computational approach w.r.t. CASA to improve overall efficiency. Within industry there are strategies viewed as either High Performance or High Throughput Computing, commonly referred to as either HPC or HTC. Neither approach is optimal for NRAO so what is proposed here is a High Efficiency Computing approach that considers all aspects of the problem including hardware configuration and performance, PI time, and costs for hardware and software development. Because the goal is broad efficiency across multiple disjoint axis there is no ultimate final goal that is being aimed at, rather there is a finite series of stages which achieve some delta improvement in efficiency with the next stage going through a review and prototype process before commencing.
https://staff.nrao.edu/wiki/bin/view/NM/HTCondor#Conversion
https://staff.nrao.edu/wiki/bin/view/NM/Slurm
DONE: Queues: We want to keep the multiple queue functionality of Torque/Moab where, for example, HERA jobs go to hera nodes and VLASS jobs go to vlass nodes. We would also like to be able to have vlasstest jobs go to the vlass nodes with a higher priority without preempting running jobs.
Slurm
- Queues are called partitions. At some level they are called partitions in Torque as well.
- Job preemtion is disabled by default
- Allows for simple priority settings in partitions with the default PriorityType=priority/basic plugin.
HERA
PartitionName=hera Nodes=herapost[001-010] Default=YES MaxTime=144000 State=UP
- User: #SBATCH -p hera
- VLASS/VLASSTEST
- Server: PartitionName=vlass Nodes=nmpost[061-090] MaxTime=144000 State=UP Priority=1000
- Server: PartitionName=vlasstest Nodes=nmpost[061-070] MaxTime=144000 State=UP
- User: #SBATCH -p vlass
- HTCondor
- HTCondor doesn't have queues or partitions like Torque/Moab or Slurm but there are still ways to do what we need.
- Constraints, Custom ClassAds, and Ranks is an option. For example, HERA nodes could set the following in their configs
- HERA = True
- STARTD_ATTRS = $(STARTD_ATTRS) HERA
START = ($(START)) && (TARGET.partition =?= "HERA")
- and users could set the following in their submit files
- Requirements = (HERA =?= True) or Requirements = (HERA == True) The differences may not be important.
+partition = "HERA"
- We could do the same for VLASS/VLASSTEST but I don't know if HTCondor can prioritize VLASS over VLASSTEST the way we do with Moab. We could also do something like this for interactive nodes and nodescheduler if we end up using that.
- VLASS = True
- VLASSTEST = True
- STARTD_ATTRS = $(STARTD_ATTRS) VLASS VLASSTEST
START = ($(START)) && (TARGET.partition =?= "VLASS")
- and users could set the following in their submit files
- requirements = (VLASS =?= True) or requirements = (VLASSTEST =?= True)
+partition = "VLASS" or +partition = "VLASSTEST" depending on which they want
Rank = (VLASS =?= True) + (VLASSTEST =!= True) if they want to run VLASS jobs on unused VLASSTEST nodes.
- Using separate pools for things like HERA and VLASS is an option, but may be overkill as it would require separate Central Managers.
- HTCondor does support accounting groups that may work like queues.
- Because of the design of HTCondor there isn't a central place to define the order and "queue" of nodes like there is in Torque.
- HTCondor doesn't have queues or partitions like Torque/Moab or Slurm but there are still ways to do what we need.
DONE: Access: Would like to prevent users from being able to login to nodes unless they have a proper reservation. Right now we restrict access via /etc/security/access.conf and use Torque's pam_pbssimpleauth.so to allow access for any user running a job.
- Slurm
- Has a pam_slurm.so module which does seem to work like the pam_pbssimpleauth.so module.
- Has a pam_slurm.so module which does seem to work like the pam_pbssimpleauth.so module.
- HTCondor
- How do we restrict access to condor nodes to only those users with valid jobs running?
- With the restrictions in access.conf, HTCondor can still run jobs as users like krowe2. I think this is because HTCondor doesn't use the login mechanism but just starts shells as the user.
- How do we restrict access to condor nodes to only those users with valid jobs running?
- Slurm
DONE: Ability to run jobs remotely (AWS, CHTC, OSG, etc)
- Slurm
- I don't think we will need this ability with Slurm
- HTCondor
- We have successfully tested both condor_annex to AWS, and flocking to CHTC.
- Slurm
DONE: Cgroups: We will need protection like what cgroups provide so that jobs can’t impact other jobs on the same node.
- Slurm
- /etc/slurm/cgroup.conf
- HTCondor
- Set CGROUP_MEMORY_LIMIT_POLICY = hard in /etc/condor/config.d/99-nrao on the execute nodes.
- Set CGROUP_MEMORY_LIMIT_POLICY = hard in /etc/condor/config.d/99-nrao on the execute nodes.
- Slurm
DONE: Submit hosts: we may have several hosts that will need to be able to submit and delete jobs. (wirth, mcilroy, hamilton, etc)
- Slurm
- Slurm-20 requires systemd so hosts must be RHEL7 or later.
- HTCondor
- Slurm
DONE: Pack Jobs: Put jobs on nodes efficiently such that as many nodes as possible are left idle and available for users with large memory and/or large core-count requirements.
- Slurm
- Add SchedulerType=sched/backfill to /etc/slurm/slurm.conf on the Management Node
- HTCondor
- Add NEGOTIATOR_DEPTH_FIRST = True to /etc/condor/config.d/99-nrao on the Central Manager
- Slurm
DONE: Reservations: The ability to reserve nodes far in the future for things like CASA classes and SIW would be very helpful. It would need to prevent HTCondor from starting jobs on these nodes as reservation time approaches.
- Slurm
- scontrol create reservation starttime=now duration=5 nodes=testpost001 user=root
- scontrol create reservation starttime=2022-05-3T08:00:00 duration=21-0:0:0 nodes=nmpost[020-030] user=root reservationname=siw2022
- scontrol show res The output of this kinda sucks. Hopefully there is a better way to see all the reservations.
- HTcondor
- There isn't a reservation feature in HTCondor. Since CHTC makes use of preemption, their nodes can be removed at almost any time without adversely affecting running jobs. Sadly NRAO cannot really use preemption.
- Slurm
- DONE: Array jobs: Do we want to keep the Torque array job functionality?
- Slurm
- #SBATCH --array=0-3%2 This syntax is very similar to Torque.
- HTCondor
- To some extent, this isn't how HTCondor is ment to be used. In other extents, DAGMan and the queue command can simulate this.
- queue 100 starts 100 copies of the job
- queue from seq 10 5 30 | will launch five jobs with $(item) set to 10, 15, 20, 25, 30
- queue item in 0, 1, 2, 3 Is another example. I don't think you can do the modulus feature with queue (i.e. %)
- You can throttle DAGMan jobs https://htcondor.readthedocs.io/en/latest/users-manual/dagman-workflows.html#throttling-nodes-by-category
- To some extent, this isn't how HTCondor is ment to be used. In other extents, DAGMan and the queue command can simulate this.
- Slurm
SKIP: Interactive: The ability to assign all or part of a node to a user with shell level access (nodescheduler, qsub -I, etc), minimal granularity is per NUMA node, finer would be useful. Because Slurm and HTCondor lack the ability to implement nodescheduler, my current thought is to ditch nodescheduler and just use the interactive commands that come with Slurm and HTCondor.
- nodescheduler: Was written before I understood what qsub -I did. Had I known, I may have argued to use qsub -I instead of nodescheduler as it is much simpler, is consistent with other installations of Torque, and may have forced some users to use batch processing which is much more efficient.
- nodescheduler likes
- It's not tied to any tty so a user can login multiple times from multiple places to their reserved node without requiring something like screen, tmux, or vnc. It also means that users aren't all going through nmpost-master.
- Its creation is asynchronous. If the cluster is full you don't wait around for your reservation to start, you get an email message when it is ready.
- It's time limited (e.g. two weeks). We might be able to do the same with a queue/partition setting but could we then extend that reservation?
- We get to define the shape of a reservation (whole node, NUMA node, etc). If we just let people use qsub -I they could reserve all sorts of sizes which may be less efficient. Then again it may be more efficient. But either way I think nodescheduler it is simpler for our users.
- It's not tied to any tty so a user can login multiple times from multiple places to their reserved node without requiring something like screen, tmux, or vnc. It also means that users aren't all going through nmpost-master.
- nodescheduler: dislikes
- With Toruqe/Moab asking for a NUMA node doesn't work as I would like. Because of bugs and limitations, I still have to ask for a specific amount of memory. The whole point of asking for a NUMA node was that I didn't need to know the resources of a node ahead of time but could just ask for half of a node. Sadly, that doesn't work with Torque/Moab.
- Because of the way I maintain the cgroup for the user, with /etc/cgrules.conf, I cannot let a user have more than one nodescheduler job on the same node or it will be impossible to know which cgroup an ssh connection should use. The interactive commands (qsub -I, etc) don't have this problem.
- Slurm
- srun --pty bash This logs the user into an interactive shell on a node with defaults (1 core, 1 GB memory)
- Slurm has system-level prolog/epilog functionality that should allow nodesceduler to set /etc/cgrules.conf.
- I don't see how Slurm can reserve NUMA nodes so we may have to just reserve X tasks with Y memory.
- I don't know how to keep Slurm from giving a user multiple portions of the same host. With Moab I used naccesspolicy=uniqueuser which prevents the ambiguity of which ssh connection goes to which cgroup. I could have nodescheduler check the nodes and assign one that the user isn't currently using but this is starting to turn nodescheduler into a scheduler of its own and I think may be more complication than we want to maintain.
- One method would be to request a node, check if that
- HTCondor
- condor_submit -i This logs the user into an interactive shell on a node with defaults (1 core equivelent, 0.5 GB memory)
- I don't see how HTCondor can reserve NUMA nodes so we may have to just reserve X tasks with Y memory.
- Could run a sleep job just like we do with Torque and use condor_ssh_to_job which seems to do X11 properly. We would probably want to make gygax part of the nmpost pool.
- I don't think I need to worry about giving a user multiple portions of the same host if we are using condor_ssh_to_job.
- How can we allow a user to login to a node they have an interactive job running on via nodescheduler? With Torque or Slurm there are PAM modules but there isn't one for HTCondor.
- nodevnc
- Given the limitation of Slurm and HTCondor and that we already recommend users use VNC on their interactive nodes, why don't we just provide a nodevnc script that reserves a node (via torque, slurm or HTCondor), start a vnc server and then tells the user it is ready and how to connect to it? If someone still needs/wants just simple terminal access, then qsub -I or srun --pty bash or condor_submit -i might suffice.
- nodescheduler: Was written before I understood what qsub -I did. Had I known, I may have argued to use qsub -I instead of nodescheduler as it is much simpler, is consistent with other installations of Torque, and may have forced some users to use batch processing which is much more efficient.
MPI: We have some users that use MPI across multiple nodes. It would be nice to keep that as an option.
- Slurm
- mpich2
- PATH=${PATH}:/usr/lib64/mpich/bin salloc --ntasks=8 mpiexec mpiexec.sh
- PATH=${PATH}:/usr/lib64/mpich/bin salloc --nodes=2 mpiexec mpiexec.sh
- OpenMPI
- Use #SBATCH to request a number of tasks (cores) and then run mpiexec or mpicasa as normal.
- mpich2
- HTCondor
- Single-node MPI jobs do work in the Vanilla universe.
- Multi-node MPI jobs require the creation of a Parallel universe. How?
- Slurm
Reaper: Clean nodes of unwanted files, dirs and procs. Condor seems to handle /tmp, /var/tmp, and /dev/shm (as of 8.9.7) properly because it uses fake versions of these dirs for each job. What about errant processes?
Slurm
There is the pam_slurm_adopt.so that supposedly tracks and kills errant processes but it conflicts with systemd and therefore requires some special tweaking.
- HTCondor
- Seems to handle /tmp and /var/tmp properly because it uses fake versions of these dirs for each job.
- What about errant processes?
- Seems to handle /tmp and /var/tmp properly because it uses fake versions of these dirs for each job.
Reaper: Cancel jobs when accounts are closed.
Node priority: With Torque/Moab we can control the order in which the scheduler picks nodes. This allows us to run jobs on the faster nodes by default.
- Slurm
- The order of the nodes in PartitionName is not important. But you can set a Weight to a NodeName. Nodes with the lowest weight will be chosen first.
- HTCondor
- There isn't a simple list like pbsnodes in Torque but there is NEGOTIATOR_PRE_JOB_RANK which can be used to weight nodes by cpu, memory, etc.
- There isn't a simple list like pbsnodes in Torque but there is NEGOTIATOR_PRE_JOB_RANK which can be used to weight nodes by cpu, memory, etc.
- Slurm
While preemption can be useful in some circumstances I expect we will want it disabled for the foreseeable future.
Slurm
The default is PreemptType=preempt/none which means Slurm will not preempt jobs.
- HTCondor
- Setting a Machine Rank will cause jobs to be preempted https://htcondor-wiki.cs.wisc.edu/index.cgi/wiki?p=HowToConfigPrioritiesForUsers
- Run both Slurm and HTCondor on the same nodes
- Server Layout: When I initially installed Torque/Maui I wanted a "secret server" that users would not log in to that ran the important services and therefore would not suffer any user shinanigans like running out of memory or CPU. This is what nmpost-serv-1 is. Then I wanted a submit host that all users would use to submit jobs and if it suffered shinanigans, jobs would not be lost or otherwise interrupted. This is what nmpost-master is. This system has performed pretty well. For HTCondor, this setup seemed to map well to their ideas of execution host, submit host, and central manager. But if we want to do flocking or condor_annex, we will need both the submit host and central manager to have external IPs. This makes me question keeping the central manager and submit hosts as separate hosts.
- The central manager will need an external IP. Will it need access to Lustre?
- The submit host that can flock will need an external IP. It will need access to Lustre.
- The submit host that runs condor_annex will need an external IP. It will need access to Lustre.
https://open-confluence.nrao.edu/download/attachments/40537022/nmpost-slurm.conf?api=v2 is a proposed slurm.conf for our nmpost cluster.
Now that we have a list of requirements I think the next step is to create a step-by-step procedure document listing everything that needs to be done to migrate from Torque/Moab to HTCondor and perhaps also Slurm.