Docs - Standardize alphabetic ordering for daemon-specific Parameters
diff --git a/doc/man/man5/slurm.conf.5 b/doc/man/man5/slurm.conf.5 index af6c541..e777e89 100644 --- a/doc/man/man5/slurm.conf.5 +++ b/doc/man/man5/slurm.conf.5
@@ -5423,6 +5423,12 @@ .IP .TP +\fBconmgr_connect_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering an outbound connection attempt to be +timed out. Defaults to the value of \fBMessageTimeout\fR. +.IP + +.TP \fBconmgr_max_connections\fR=<\fIconnection_count\fR> Specify the maximum number of connections to be processed at any given time. This does not influence the maximum number of pending connections as that is @@ -5432,6 +5438,19 @@ .IP .TP +\fBconmgr_quiesce_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering quiesce to be timed out. Upon timeout, +all (non-listening) active connections will be closed to allow the quiesce to +start. Defaults to two times value of \fBMessageTimeout\fR. +.IP + +.TP +\fBconmgr_read_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering a read from a file descriptor to be +timed out. Defaults to the value of \fBMessageTimeout\fR. +.IP + +.TP \fBconmgr_threads\fR=<\fIthread_count\fR> Alias for \fBworkerpool_threads\fR. .IP @@ -5442,25 +5461,6 @@ .IP .TP -\fBconmgr_connect_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering an outbound connection attempt to be -timed out. Defaults to the value of \fBMessageTimeout\fR. -.IP - -.TP -\fBconmgr_read_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering a read from a file descriptor to be -timed out. Defaults to the value of \fBMessageTimeout\fR. -.IP - -.TP -\fBconmgr_quiesce_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering quiesce to be timed out. Upon timeout, -all (non-listening) active connections will be closed to allow the quiesce to -start. Defaults to two times value of \fBMessageTimeout\fR. -.IP - -.TP \fBconmgr_wait_write_delay\fR=<\fIseconds\fR> When waiting for kernel to flush outgoing buffer, poll kernel for changes every <\fIseconds\fR> for changes. Defaults to the value of \fBMessageTimeout\fR. @@ -5503,6 +5503,14 @@ .IP .TP +\fBenable_stepmgr\fR +Enable slurmstepd step management system wide. This enables job steps to be +managed by a single extern slurmstepd associated with the job to manage steps. +This is beneficial for jobs that submit many steps inside their allocations. +\fBPrologFlags=contain\fR must be set. +.IP + +.TP \fBidle_on_node_suspend\fR Mark nodes as idle, regardless of current state, when suspending nodes with \fBSuspendProgram\fR so that nodes will be eligible to be resumed at a later @@ -5510,6 +5518,30 @@ .IP .TP +\fBmax_dbd_msg_action\fR +Action used once MaxDBDMsgs is reached, options are 'discard' (default) and 'exit'. + +When 'discard' is specified and MaxDBDMsgs is reached we start by purging +pending messages of types Step start and complete, and it reaches MaxDBDMsgs +again Job start messages are purged. Job completes and node state changes +continue to consume the empty space created from the purgings until MaxDBDMsgs +is reached again at which no new message is tracked creating data loss and +potentially runaway jobs. + +When 'exit' is specified and MaxDBDMsgs is reached the slurmctld will exit +instead of discarding any messages. It will be impossible to start the +slurmctld with this option where the slurmdbd is down and the slurmctld is +tracking more than MaxDBDMsgs. +.IP + +.TP +\fBmax_powered_nodes\fR +The max number of powered up nodes across the cluster. Once this is reached, +jobs requesting additional nodes will not start, and "scontrol power up +<nodes>" will fail. +.IP + +.TP \fBnode_reg_mem_percent\fR=\# Percentage of memory a node is allowed to register with without being marked as invalid with low memory. Default is 100. For State=CLOUD nodes, the default is @@ -5543,30 +5575,6 @@ .IP .TP -\fBmax_powered_nodes\fR -The max number of powered up nodes across the cluster. Once this is reached, -jobs requesting additional nodes will not start, and "scontrol power up -<nodes>" will fail. -.IP - -.TP -\fBmax_dbd_msg_action\fR -Action used once MaxDBDMsgs is reached, options are 'discard' (default) and 'exit'. - -When 'discard' is specified and MaxDBDMsgs is reached we start by purging -pending messages of types Step start and complete, and it reaches MaxDBDMsgs -again Job start messages are purged. Job completes and node state changes -continue to consume the empty space created from the purgings until MaxDBDMsgs -is reached again at which no new message is tracked creating data loss and -potentially runaway jobs. - -When 'exit' is specified and MaxDBDMsgs is reached the slurmctld will exit -instead of discarding any messages. It will be impossible to start the -slurmctld with this option where the slurmdbd is down and the slurmctld is -tracking more than MaxDBDMsgs. -.IP - -.TP \fBreboot_from_controller\fR Run the \fBRebootProgram\fR from the controller instead of on the slurmds. The RebootProgram will be passed a comma\-separated list of nodes to reboot as the @@ -5632,19 +5640,6 @@ .IP .TP -\fBenable_stepmgr\fR -Enable slurmstepd step management system wide. This enables job steps to be -managed by a single extern slurmstepd associated with the job to manage steps. -This is beneficial for jobs that submit many steps inside their allocations. -\fBPrologFlags=contain\fR must be set. -.IP - -.TP -\fBuser_resv_delete\fR -Allow any user able to run in a reservation to delete it. -.IP - -.TP \fBthreadpool\fR=<\fIenabled|disabled\fR> Enable or disable using the threadpool in slurmctld to avoid overhead of allocating new process threads by preserving previously created threads. @@ -5666,6 +5661,11 @@ .IP .TP +\fBuser_resv_delete\fR +Allow any user able to run in a reservation to delete it. +.IP + +.TP \fBvalidate_nodeaddr_threads\fR= During startup, slurmctld looks up the address for each compute node in the system. On large systems this can cause considerable delay, this option permits @@ -5909,6 +5909,12 @@ .IP .TP +\fBconmgr_connect_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering an outbound connection attempt to be +timed out. Defaults to the value of \fBMessageTimeout\fR. +.IP + +.TP \fBconmgr_max_connections\fR=<\fIconnection_count\fR> Specify the maximum number of connections to be processed at any given time. This does not influence the maximum number of pending connections as that is @@ -5918,6 +5924,19 @@ .IP .TP +\fBconmgr_quiesce_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering quiesce to be timed out. Upon timeout, +all (non-listening) active connections will be closed to allow the quiesce to +start. Defaults to two times value of \fBMessageTimeout\fR. +.IP + +.TP +\fBconmgr_read_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering a read from a file descriptor to be +timed out. Defaults to the value of \fBMessageTimeout\fR. +.IP + +.TP \fBconmgr_threads\fR=<\fIthread_count\fR> Alias for \fBworkerpool_threads\fR. .IP @@ -5928,25 +5947,6 @@ .IP .TP -\fBconmgr_connect_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering an outbound connection attempt to be -timed out. Defaults to the value of \fBMessageTimeout\fR. -.IP - -.TP -\fBconmgr_read_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering a read from a file descriptor to be -timed out. Defaults to the value of \fBMessageTimeout\fR. -.IP - -.TP -\fBconmgr_quiesce_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering quiesce to be timed out. Upon timeout, -all (non-listening) active connections will be closed to allow the quiesce to -start. Defaults to two times value of \fBMessageTimeout\fR. -.IP - -.TP \fBconmgr_wait_write_delay\fR=<\fIseconds\fR> When waiting for kernel to flush outgoing buffer, poll kernel for changes every <\fIseconds\fR> for changes. Defaults to the value of \fBMessageTimeout\fR. @@ -5959,6 +5959,13 @@ .IP .TP +\fBcontain_spank\fR +If set and a namespace plugin is specified, the spank_user(), +spank_task_post_fork() and spank_task_exit() calls will be run inside the job's +namespace. +.IP + +.TP \fBl3cache_as_socket\fR Use the hwloc l3cache as the socket count. Can be useful on certain processors where the socket level is too coarse, and the l3cache may provide better @@ -6146,6 +6153,12 @@ .IP .RS .TP +\fBconmgr_connect_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering an outbound connection attempt to be +timed out. Defaults to the value of \fBMessageTimeout\fR. +.IP + +.TP \fBconmgr_max_connections\fR=<\fIconnection_count\fR> Specify the maximum number of connections to be processed at any given time. This does not influence the maximum number of pending connections as that is @@ -6159,19 +6172,10 @@ .IP .TP -\fBconmgr_threads\fR=<\fIthread_count\fR> -Alias for \fBworkerpool_threads\fR. -.IP - -.TP -\fBconmgr_use_poll\fR -Use \fIpoll\fR(2) instead of \fIepoll\fR(7) for monitoring file descriptors. -.IP - -.TP -\fBconmgr_connect_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering an outbound connection attempt to be -timed out. Defaults to the value of \fBMessageTimeout\fR. +\fBconmgr_quiesce_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering quiesce to be timed out. Upon timeout, +all (non-listening) active connections will be closed to allow the quiesce to +start. Defaults to two times the value of \fBMessageTimeout\fR. .IP .TP @@ -6181,10 +6185,13 @@ .IP .TP -\fBconmgr_quiesce_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering quiesce to be timed out. Upon timeout, -all (non-listening) active connections will be closed to allow the quiesce to -start. Defaults to two times the value of \fBMessageTimeout\fR. +\fBconmgr_threads\fR=<\fIthread_count\fR> +Alias for \fBworkerpool_threads\fR. +.IP + +.TP +\fBconmgr_use_poll\fR +Use \fIpoll\fR(2) instead of \fIepoll\fR(7) for monitoring file descriptors. .IP .TP @@ -6239,6 +6246,12 @@ .IP .RS .TP +\fBconmgr_connect_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering an outbound connection attempt to be +timed out. Defaults to the value of \fBMessageTimeout\fR. +.IP + +.TP \fBconmgr_max_connections\fR=<\fIconnection_count\fR> Specify the maximum number of connections to be processed at any given time. This does not influence the maximum number of pending connections as that is @@ -6248,6 +6261,19 @@ .IP .TP +\fBconmgr_quiesce_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering quiesce to be timed out. Upon timeout, +all (non-listening) active connections will be closed to allow the quiesce to +start. Defaults to two times value of \fBMessageTimeout\fR. +.IP + +.TP +\fBconmgr_read_timeout\fR=<\fIseconds\fR> +Wait <\fIseconds\fR> before considering a read from a file descriptor to be +timed out. Defaults to the value of \fBMessageTimeout\fR. +.IP + +.TP \fBconmgr_threads\fR=<\fIthread_count\fR> Alias for \fBworkerpool_threads\fR. .IP @@ -6258,25 +6284,6 @@ .IP .TP -\fBconmgr_connect_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering an outbound connection attempt to be -timed out. Defaults to the value of \fBMessageTimeout\fR. -.IP - -.TP -\fBconmgr_read_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering a read from a file descriptor to be -timed out. Defaults to the value of \fBMessageTimeout\fR. -.IP - -.TP -\fBconmgr_quiesce_timeout\fR=<\fIseconds\fR> -Wait <\fIseconds\fR> before considering quiesce to be timed out. Upon timeout, -all (non-listening) active connections will be closed to allow the quiesce to -start. Defaults to two times value of \fBMessageTimeout\fR. -.IP - -.TP \fBconmgr_wait_write_delay\fR=<\fIseconds\fR> When waiting for kernel to flush outgoing buffer, poll kernel for changes every <\fIseconds\fR> for changes. Defaults to the value of \fBMessageTimeout\fR.