Skip to end of metadata
Go to start of metadata


Objective

This is to introduce the node ramp down feature, which basically releases no longer needed sister nodes/vnodes early from running jobs.

Interface 1: New command: 'pbs_release_nodes'

  • Visibility: Public
  • Change Control: Stable
  • Synopsis: Release a specified set of sister nodes or vnodes, or all sister nodes or vnodes assigned to the specified running job. The nodes released will then be made available for scheduling other jobs. 
  • Permission: Only job owner, admin, PBS manager, or PBS operator will be allowed to perform the release nodes action.
  • Details:
    • Two types of  actions:
      1.  Release a particular set of sister nodes from a job:
        Syntax:   pbs_release_nodes -j <job_identifier> <host1_or_vnode1> [<host2_or_vnode2> [<host3_or_vnode3>] ...] ...

        • The 'host*_or_vnode*' argument is any of the sister nodes/vnodes that appear in the exec_vnode attribute of a running job. Example:
          % qsub job.scr
          241.borg
          % qstat  241 | grep "exec|Resource_List|select"

          exec_host = borg[0]/0*0+federer/0*0+lendl/0*2
          exec_vnode = (borg[0]:mem=1048576kb:ncpus=1+borg[1]:mem=1048576kb:ncpus=1+borg[2]:ncpus=1)+(federer:mem=1048576kb:ncpus=1+federer[0]:mem=1048576k:ncpus=1+federer[1]:ncpus=1)+(lendl:ncpus=2:mem=2097152kb)
          Resource_List.mem = 6gb
          Resource_List.ncpus = 8
          Resource_List.nodect = 3
          Resource_List.place = scatter
          Resource_List.select = ncpus=3:mem=2gb+ncpus=3:mem=2gb+ncpus=2:mem=2gb
          schedselect = 1:ncpus=3:mem=2gb+1:ncpus=3:mem=2gb+1:ncpus=2:mem=2gb


          %  pbs_release_nodes -j 241 federer[1] lendl

          % qstat  241 | grep "exec|Resource_List|select"

          exec_host = borg[0]/0*0+federer/0*0 <- no lendl as all assigned vnodes in lendl have been cleared.
          exec_vnode = (borg[0]:mem=1048576kb:ncpus=1+borg[1]:mem=1048576kb:ncpus=1+borg[2]:ncpus=1)+(federer:mem=1048576kb:ncpus=1+federer[0]:mem=1048576kb:ncpus=1) <- federer[1] and lendl removed.

          Resource_List.mem = 4194304kb <- minus 2gb (from lendl)
          Resource_List.ncpus = 5 <- minus 3 cpus (1 from federer[1] and 2 from lendl)
          Resource_List.nodect = 2 <- minus 1 chunk (when lendl was taken out, its entire chunk assignment disappeared)
          Resource_List.place = scatter
          schedselect = 1:mem=2097152kb:ncpus=3+1:mem=2097152kb:ncpus=2


      2. Release all sister nodes from a job:
        Syntax:   pbs_release_nodes -j <job_identifier> -a
        • Example:
          % pbs_release_nodes -j 241 -a
          % qstat -f 241

          exec_host = borg[0]/0*0
          exec_vnode = (borg[0]:mem=1048576kb:ncpus=1)+borg[1]:mem=1048576kb:ncpus=1+borg[2]:ncpus=1)
          Resource_List.mem = 2097152kb
          Resource_List.ncpus =  3
          Resource_List.nodect = 1
          Resource_List.place = scatter
          schedselect = 1:mem=2097152kb:ncpus=3


    • After issuing pbs_release_nodes, a running job's $PBS_NODEFILE content will  no longer show the released nodes.
      Example:

      % qsub -l select=2:ncpus=1:mem=1gb -l place=scatter -I
      qsub: waiting for job 247.borg.pbspro.com to start
      qsub: job 247.borg.pbspro.com ready

      % cat $PBS_NODEFILE
      borg.pbspro.com
      federer.pbspro.com
      %  pbs_release_nodes -j 247 federer
      % cat $PBS_NODEFILE
      borg.pbspro.com

    • The server will continue to hold on to the job on a released node, until receiving a confirmation that the job has been cleaned up from the node
    •  The PBS licenses will be updated accordingly once the job has been completely taken out of the released node.
    • When a node is released, it reports to the mother superior (MS) its resources_used* values for the job as the final action. That released node would no longer update the resources_used values for that job since it's no longer part of the job. But MS will hold onto the data, and will be added during final aggregation of resources_used values when job exits.
    • pbs_release_nodes is not currently supported with nodes/vnodes that are tied to Cray XC systems,  as the ALPS reservation cannot be modified right now. 
    • pbs_release_nodes is also not supported with nodes/vnodes managed by cpuset moms, given that partial release of vnodes may result in leftover cpusets.
    • If cgroups support is enabled, and pbs_release_nodes is called to release some of the vnodes but not all the vnodes from the same mom host, resources on those vnodes that are part of a cgroup would not get automatically released.
    •  Error reporting

        •  pbs_release_nodes will report an error if any of the nodes specified are managed by a mother superior mom.

          Example:

          % pbs_release_nodes -j 241 borg[0]

            pbs_release_nodes: Can't free 'borg[0]' since it's on an MS host

        • pbs_release_nodes will report an error if executed by a non-admin, non-manager, non-operator, or non-job owner user:

          Example:

          %  pbs_release_nodes -j 248 federer
          pbs_release_nodes: Unauthorized Request

        • pbs_release_nodes will report an error if the vnode being released is not part of the job:
          Example:

          %  pbs_release_nodes -j 249 lendl
          pbs_release_nodes: these nodes are not part of the job: lendl

        • pbs_release_nodes will report an error if issued on a job that is not in a running state:
          Example:

          % pbs_release_nodes -j 251 lendl
          pbs_release_nodes: Request invalid for state of job

        • pbs_release_nodes will report an error if both the '-a' option and a list of nodes/vnodes are specified in the command line.

          Example:

          % pbs_release_nodes -j 252 -a federer
          usage: pbs_release_nodes [-j job_identifier] host_or_vnode1 host_or_vnode2 ...
          usage: pbs_release_nodes [-j job_identifier] -a
          pbs_release_nodes --version

        • pbs_release_nodes will report an error message and exit if at least one of the hosts or vnodes specified is a Cray XC series node. The following message is returned:

          Example:

          % pbs_release_node -j 253 cray_node

          "pbs_release_nodes: not currently supported on Cray XC systems nodes: <cray_node>"

        • pbs_release_nodes will report an error message and exit if at least one of the hosts or vnodes specified is managed by a pbs_mom running cpusets (i.e. resources_available.arch='linux_cpuset'). The following message is returned:

          Example:

          % pbs_release_node -j 253 cray_node

          "pbs_release_nodes: not currently supported on nodes whose resources are part of a cpuset: <cpuset_node>"


    • At every successful pbs_release_nodes call, qstat will show the updated exec_host, exec_vnode, Resource_List* values.

    • When releasing vnodes, if all vnodes assigned coming from the same mom host have been released, then the job would be completely removed from that mom host. This will result in 1) execjob_epilogue hook script (if it exists) to execute, 2) job processes are killed on that mom host, 3) any job-specific specific files including job temporary directories are removed, and 4) cpusets and Cray alps reservations on that mom host are cleared. Note that the execjob_end hook will not execute on the host in this case.

    • If one (or more) but not all the vnodes from a mom host assigned to the job have been released (partial release of vnodes), then job does not get removed from the mom host yet. If those released vnodes have been configured to be shared, then they can be reassigned to other jobs.

    • API:

      NAME
      pbs_relnodesjob - release a set of sister nodes or vnodes,
      or all sister nodes or vnodes assigned to the specified PBS
      batch job.

      SYNOPSIS
      #include <pbs_error.h>
      #include <pbs_ifl.h>

      int pbs_relnodesjob(int connect, char *job_id, char *node_list,
      char *extend)

      DESCRIPTION
      Issue a batch request to release sister vnodes from a batch job.

      A RelnodesJob batch request is generated and sent to the server over
      the connection specified by connect which is the return value of
      pbs_connect().

      The argument, job_id, identifies the job to release nodes or vnodes
      from; it is specified in the form:
      sequence_number.server

      The parameter, node_list, is a plus ('+') separated list of vnode names
      whose whose parent mom is a sister mom. If node_list is NULL, then
      this refers to all the sister vnodes assigned to the job.

      The parameter, extend, is reserved for implementation-defined exten-
      sions.

Interface 2: New job attribute 'release_nodes_on_stageout'

  • Visibility: Public
  • Change Control: Stable
  • Value: 'true' or 'false'
  • Synopsis: When set to 'true', this will do an equivalent of 'pbs_release_nodes -a' for releasing all the sister nodes when stageout operation begins.
  • Example:
    %  qsub -W stageout=my_stageout@federer:my_stageout.out -W release_nodes_on_stageout=true job.scr

  • This can also be specified in the server attribute 'default_qsub_arguments' to allow all jobs to be submitted with release_nodes_stageout set by default.

  • This attribute can also be set in a queuejob, modifyjob hook, and the Python type is boolean with valid values 'True' or 'False'.

Example:
 # cat qjob.py
import pbs
e=pbs.event()
e.job.release_nodes_on_stageout = True
# qmgr -c "create hook qjob event=queuejob"
# qmgr -c "import hook application/x-python default qjob.py"
% qsub job.scr
23.borg
% qstat -f 23
...
release_nodes_on_stageout = True

Interface 3: New server accounting record: 'u' for update record

  • Visibility: Public
  • Change Control: Stable
  • Synopsis: For every release nodes action, there'll be an accounting_logs record written, which is called the

    'u' (for update) record.

  • Details: The 'u' record represents a just concluded phase of the job, which  consists of a set of resources assigned to the job ( exec_vnode, exec_host, Resource_List items), and amount of resources used (resources_used) during that phase of the job.

  • Example:

% qsub -l select=3:ncpus=1:mem=1gb job.scr

242.borg

% qstat -f 242 | egrep "exec|Resource_List|select"

exec_host = borg/0+federer/0+lendl/0

exec_vnode = (borg[0]:ncpus=1:mem=1048576kb)+(federer:ncpus=1:mem=1048576kb)+(lendl:ncpus=1:mem=1048576kb)

Resource_List.mem = 3gb

Resource_List.ncpus = 3

Resource_List.nodect = 3

Resource_List.place = scatter

Resource_List.select = 3:ncpus=1:mem=1gb

schedselect = 3:ncpus=1:mem=1gb

% pbs_release_nodes -j 241 lendl

Accounting logs show:

# tail -f /var/spool/PBS/server_priv/accounting/201701231

23/2017 18:53:24;u;242.borg.user=bayucan group=users project=_pbs_project_default jobname=STDIN queue=workq ctime=1485215572 qtime=1485215572 etime=1485215572 start=1485215572  session=7503 run_count=1 exec_host=borg/0+federer/0+lendl/0 exec_vnode=(borg[0]:ncpus=1:mem=1048576kb)+(federer:ncpus=1:mem=1048576kb)+(lendl:ncpus=1:mem=1048576kb) Resource_List.mem=3gb Resource_List.ncpus=3 Resource_List.nodect=3 Resource_List.place=scatter Resource_List.select=3:ncpus=1:mem=1gb resources_used.cpupercent=5 resources_used.cput=00:04:35 resources_used.mem=4288kb resources_used.ncpus=3 resources_used.vmem=42928kb resources_used.walltime=00:00:26

Another pbs_release_nodes call yield:

% pbs_release_nodes -j 241 federer

# tail -f /var/spool/PBS/server_priv/accounting/201701231

01/23/2017 18:59:35;u;242.borg;user=bayucan group=users project=_pbs_project_default jobname=STDIN queue=workq ctime=1485215572 qtime=1485215572 etime=1485215572 start=1485215572 session=7773 run_count=1  exec_host=borg/0+federer/0 exec_vnode=(borg[0]:ncpus=1:mem=1048576kb)+(federer:ncpus=1:mem=1048576kb) Resource_List.mem=2097152kb Resource_List.ncpus=2 Resource_List.nodect=2 Resource_List.place=scatter Resource_List.select=1:ncpus=1:mem=1048576kb+1:ncpus=1:mem=1048576kb resources_used.cpupercent=3 resources_used.cput=00:03:35 resources_used.mem=2048kb resources_used.ncpus=2 resources_used.vmem=32928kb resources_used.walltime=00:00:26 

Interface 4: New server accounting record: 'c' for continue record:

  • Visibility: Public
  • Change Control: Stable
  • Synopsis: The 'c' accounting record will show  the next assigned exec_vnode, exec_host, Resource_List. along with the job attributes in the new/next phase of the job.

    Given the following example:                                                                                                                                                                                  % qsub -l select=3:ncpus=1:mem=1gb job.scr

    242.borg

    % qstat -f 242 | egrep "exec|Resource_List|select"

    exec_host = borg/0+federer/0+lendl/0

    exec_vnode = (borg[0]:ncpus=1:mem=1048576kb)+(federer:ncpus=1:mem=1048576kb)+(lendl:ncpus=1:mem=1048576kb)

    Resource_List.mem = 3gb

    Resource_List.ncpus = 3

    Resource_List.nodect = 3

    Resource_List.place = scatter

    Resource_List.select = 3:ncpus=1:mem=1gb

    schedselect = 3:ncpus=1:mem=1gb

    % pbs_release_nodes -j 241 lendl

    Accounting logs show:

    # tail -f /var/spool/PBS/server_priv/accounting/201701231

    23/2017 18:53:24;u;242.borg.user=bayucan group=users project=_pbs_project_default jobname=STDIN queue=workq ctime=1485215572 qtime=1485215572 etime=1485215572 start=1485215572  session=7503 run_count=1 exec_host=borg/0+federer/0+lendl/0 exec_vnode=(borg[0]:ncpus=1:mem=1048576kb)+(federer:ncpus=1:mem=1048576kb)+(lendl:ncpus=1:mem=1048576kb) Resource_List.mem=3gb Resource_List.ncpus=3 Resource_List.nodect=3 Resource_List.place=scatter Resource_List.select=3:ncpus=1:mem=1gb resources_used.cpupercent=5 resources_used.cput=00:04:35 resources_used.mem=4288kb resources_used.ncpus=3 resources_used.vmem=42928kb resources_used.walltime=00:00:26

    % pbs_release_nodes -j 241 lendl

    Accounting logs show (lendl vnode assignment gone):

    # tail -f /var/spool/PBS/server_priv/accounting/201701231

    23/2017 18:53:24;c;242.borg.user=bayucan group=users project=_pbs_project_default jobname=STDIN queue=workq ctime=1485215572 qtime=1485215572 etime=1485215572 start=1485215572  session=7503 run_count=1 exec_host=borg/0+federer/0 exec_vnode=(borg[0]:ncpus=1:mem=1048576kb)+(federer:ncpus=1:mem=1048576kb) Resource_List.mem=2097152kb Resource_List.ncpus=2 Resource_List.nodect=2 Resource_List.place=scatter updated_Resource_List.select=1:ncpus=1:mem=1048576kb+1:ncpus=1:mem=1048576kb resources_used_incr.cpupercent=5 r

    Another pbs_release_nodes call yield 'federer' vnode assignment gone::

    % pbs_release_nodes -j 241 federer

    # tail -f /var/spool/PBS/server_priv/accounting/201701231

    01/23/2017 18:59:35;c;242.borg;user=bayucan group=users project=_pbs_project_default jobname=STDIN queue=workq cctime=1485215572 qtime=1485215572 etime=1485215572 start=1485215572 session=7773 run_count=1  exec_host=borg/0 exec_vnode=(borg[0]:ncpus=1:mem=1048576kb) Resource_List.mem=1048576kb Resource_List.ncpus=1 Resource_List.nodect=1 Resource_List.place=scatter Resource_List.select=1:ncpus=1:mem=1048576kb

Interface 5: New server accounting record: 'e' (end) for end of job record for a phased job

  • Visibility: Public
  • Change Control: Stable
  • Synopsis: The 'e' accounting record will show the resources assigned to the job ( exec_vnode, exec_host, Resource_List items), and amount of resources used (resources_used) during that last phase of the job.
  • Details: It will be up to the log parser to take all the 'u' records and 'e' record of the job and either sum up the resources_used.* values (e.g. resources_used.walltime) or to average them out whichever makes sense (e.g. resources_used.ncpus). Note that the regular 'E' (end) accounting record will continue to be generated for a job, whether it has released nodes or not, showing the job's values in total at the end of the job:

Interface 6: Additions to log messages

  • Visibility: Public
  • Change Control: Stable
  • Details:
    • Special mom_logs messages:
      • A pbs_release_nodes request causes the server to send a job update to the mother superior (MS)  of the job. The
        MS in turn looks into the list of nodes being removed. If it's the last node from the same host, MS sends a new DELETE_JOB2 request to that owning sister mom. Upon receiving this request, the sister mom goes and kills job processes on the node, and sends back to the mother superior the summary
        accounting information for the job on that node. Mom_logs will show the following DEBUG messages:

        sister mom_logs: "DELETE_JOB2 received"

        Mother superior log: "<reporting_sister_host>;cput=YY mem=ZZ"

    • Special server_logs messages:

      • When a job has been completely removed from an early released vnode, the following DEBUG2 message will be shown:

        "clearing job <job-id> from node <vnode-name>

        "Node<sister-mom-hostname>;deallocating 1 cpus from job <job-id>

Interface 7: New server attribute 'show_hidden_attribs'

  • Visibility: Private
  • Change Control: Experimental
  • Value: 'true' or 'false'
  • Synopsis: When set to 'true', this allows qstat -f to also show values of internal attributes created by the server to implement the node ramp down feature. Some example internal job attributes that may show its value are exec_vnode_orig, exec_vnode_acct, exec_vnode_deallocated, exec_host_orig, exec_host_acct, Resource_List_orig, Resource_List_acct.

  • Note: This attribute is provided as an aid to debugging PBS.


  • No labels