SLURM 19 discontinued the use of --cpu_bind (and changed it to
--cpu-bind). There's no easy way to test at run time which one is
accepted, so set the environment variable SLURM_CPU_BIND to "none",
which should do the same thing as the srun CLI parameter.
Signed-off-by: Jordan Hayes <jhayes@ucr.edu>
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
* When we moved to allowing dual rml/oob transports, we added a bunch of
stuff that is no longer needed. Remove it so as to simplify the
messaging system.
* Fix the routed/radix component so it correctly returns the parent's
vpid
Signed-off-by: Ralph Castain <rhc@pmix.org>
We've been fighting the battle of trying to create a regex generator and
parser that can handle arbitrary hostname schemes - without long-term
success. The worst of it is that there is no way of checking to see if
the computed regex is correct short of parsing it and doing a
character-by-character comparison with the original string. Ugh...there
has to be a better solution.
One option is to investigate using 3rd-party regex libraries as
those are coming from communities whose sole focus is resolving that
problem. However, someone would need to spend the time to investigate
it, and we'd have to find a license-friendly implementation.
Another option is to quit beating our heads against the wall and just
compress the information. It won't be as much of a reduction, but we
also won't keep hitting scenarios where things break. In this case, it
seems that "perfection" is definitely the enemy of "good enough".
This PR implements the compression option while retaining the
possibility of people adding regex-generating components. The
compression code used in ORTE is consolidated into the opal/compress
framework. That framework currently held bzip and gzip components for
use in compressing checkpoint files - since we no longer support C/R, I
have .opal_ignore'd those components.
However, I have left the original framework APIs alone in case someone
ever decides to redo C/R. The APIs of interest here are added to the
framework - specifically, the "compress_block" and "decompress_block"
functions. I then moved the ORTE zlib compression code into a new
component in this framework.
Unfortunately, the framework currently is a single-select one - i.e.,
only one active component at a time. Since I .opal_ignore'd the other
two and made the priority of zlib high, this isn't a problem. However,
if someone wants to re-enable bzip/gzip or add another component, they
might need to transition opal/compress to a multi-select framework.
Included changes:
* Consolidate the compression code into the opal/compress framework
* Move the ORTE zlib compression code into a new opal/compress/zlib
component
* Ignore the bzip and gzip components in opal/compress framework
* Add a "compress_base_limit" MCA param to set the threshold above which
we compress data - defaults to 4096 bytes
* Delete stale brucks and rcd components from orte/grpcomm framework
* Delete the orte/regx framework
* Update the launch system to use opal/compress instead of string regex
* Provide a default module if no zlib is available
* Fix some misc multi-node issues
* Properly generate the nidmap in response to a "connection warmup"
message so the remote daemon knows the children it needs to launch.
* Remove stale references to orte_node_regex
* opal_byte_object_t's are not OPAL objects - properly release allocated
memory.
* Set the topology
* Currently only handling homogeneous case
* Update the compress framework files to conform
* Consolidate open/close into one "frame" file. Ensure we open/close the
framework
Signed-off-by: Ralph Castain <rhc@pmix.org>
When there is no alias for a given node, do not set the
ORTE_NODE_ALIAS attribute to an empty string any more.
Thanks Erico for reporting this issue.
Thanks Ralph for the guidance.
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
An incorrectly named variable caused all pml variables to disappear
from ompi_info. This commit fixes the typo. We may add some logic into
the MCA base to catch these sorts of things in the future.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
While it may be faster to reverse the order of the assignment loops, it also results in the wrong answer
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Reseting the `ORTE_NODE_FLAG_MAPPED` flag after hosts filtering, this
flag is used subsequently and can be affect to the node mapping logic
Signed-off-by: Boris Karasev <karasev.b@gmail.com>
This is a point-in-time update that includes support for several new PMIx features, mostly focused on debuggers and "instant on":
* initial prototype support for PMIx-based debuggers. For the moment, this is restricted to using the DVM. Supports direct launch of apps under debugger control, and indirect launch using prun as the intermediate launcher. Includes ability for debuggers to control the environment of both the launcher and the spawned app procs. Work continues on completing support for indirect launch
* IO forwarding for tools. Output of apps launched under tool control is directed to the tool and output there - includes support for XML formatting and output to files. Stdin can be forwarded from the tool to apps, but this hasn't been implemented in ORTE yet.
* Fabric integration for "instant on". Enable collection of network "blobs" to be delivered to network libraries on compute nodes prior to local proc spawn. Infrastructure is in place - implementation will come later.
* Harvesting and forwarding of envars. Enable network plugins to harvest envars and include them in the launch msg for setting the environment prior to local proc spawn. Currently, only OmniPath is supported. PMIx MCA params control which envars are included, and also allows envars to be excluded.
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Since output-filename has been moved to a per-job attribute,
remove the orte_output_filename global variable, and stop passing
this option to orted.
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
Sync command line output for Slurm with RSH launcher.
Currently Slurm launch cmdline will only be visible in debug mode, while for RSH
it is enabled always.
cmdline makes sense for troubleshooting and should be enabled.
Signed-off-by: Artem Polyakov <artpol84@gmail.com>
Fix a problem in packing/unpacking job updates. There remains a race condition that causes messages to attempt to be sent to the second new daemon before it is completely ready. Not entirely sure where it is coming from.
Refs #4665
Rebase to master. Reset orte_nidmap_communicated if hosts are added. Check for duplicate hostnames in an add_host command. Turn off tree_spawn for dynamic launch of additional daemons.
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Handle the need for different regex generator/parsers by moving the
orte/util/nidmap and orte/util/regex code into a new "regx" framework.
Use the original code to complete a "fwd" component, and create a
scaffold for IBM's "reverse" component.
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Now that the daemon calls remote_spawn itself, there is no longer
a need for the "tree_spawn" command nor the associated command
processing code since the HNP is no longer sending a tree-spawn
message to the orted.
Thanks Ralph for the guidance !
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
When the node regex is too long to be sent on the command line,
retrieve it first from the parent, and then spawn the remote orted
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
This parameter can be used to set the node regex max length that can
be passed to the orted command line.
For testing purpose, it can be set to zero in order to force the node regex
being retrieved by orted from its parent.
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
since open-mpi/ompi@8f496b01b7
rsh_wait_daemon is invoked with an orte_wait_tracker_t *,
that must be used to reach the orte_plm_rsh_caddy_t *.
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
Somehow, the code for passing a daemon's parent was accidentally removed, thus breaking the tree-spawn callback sequence and causing all daemons to phone directly home. Note that this is noticeably slower than no-tree-spawn for small clusters where directly ssh launch of the child daemons from the HNP doesn't overload the available file descriptors.
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Change the determination of #spawn threads to be done on basis of #local procs in first job being spawned. Someone can look at an optimization that handles subsequent dynamic spawns that might be larger in size.
Leave the threads running, but blocked, for the life of the daemon, and use them to harvest the local procs as they terminate. This helps short-lived jobs in particular.
Add MCA params to set:
* max number of spawn threads (default: 4)
* set a specific number of spawn threads (default: -1, indicating no set number)
* cutoff - minimum number of local procs before using spawn threads (default: 32)
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Still in the "needs to be done" category:
* mapping/ranking/binding options aren't correctly supported
* if the DVM encounters some errors (e.g., not enough resources for the job), the resulting error is globally set and impacts any subsequent job submission
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
* Resolves#3705
* Components should link against the project level library to better
support `dlopen` with `RTLD_LOCAL`.
* Extend the `mca_FRAMEWORK_COMPONENT_la_LIBADD` in the `Makefile.am`
with the appropriate project level library:
```
MCA components in ompi/
$(top_builddir)/ompi/lib@OMPI_LIBMPI_NAME@.la
MCA components in orte/
$(top_builddir)/orte/lib@ORTE_LIB_PREFIX@open-rte.la
MCA components in opal/
$(top_builddir)/opal/lib@OPAL_LIB_PREFIX@open-pal.la
MCA components in oshmem/
$(top_builddir)/oshmem/liboshmem.la"
```
Note: The changes in this commit were automated by the script in
the commit that proceeds it with the `libadd_mca_comp_update.py`
script. Some components were not included in this change because
they are statically built only.
Signed-off-by: Joshua Hursey <jhursey@us.ibm.com>
Update to support passing of HWLOC shmem topology to client procs
Update use of distance API per @bgoglin
Have the openib component lookup its object in the distance matrix
Bring usnic up-to-date
Restore binding for hwloc2
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
passed to make it all flow thru the opal/pmix "put/get" operations. Update the PMIx code to latest master to pickup some required behaviors.
Remove the no-longer-required get_contact_info and set_contact_info from the RML layer.
Add an MCA param to allow the ofi/rml component to route messages if desired. This is mainly for experimentation at this point as we aren't sure if routing wi
ll be beneficial at large scales. Leave it "off" by default.
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Remove the opal_ignore from the RML/OFI component, but disable that component unless the user specifically requests it via the "rml_ofi_desired=1" MCA param. This will let us test compile in various environments without interfering with operations while we continue to debug
Fix an error when computing the number of infos during server init
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Start updating the various mappers to the new procedure. Remove the stale lama component as it is now very out-of-date. Bring round_robin and PPR online, and modify the mindist component (but cannot test/debug it).
Remove unneeded test
Fix memory corruption by re-initializing variable to NULL in loop
Resolve the race condition identified by @ggouaillardet by resetting the
mapped flag within the same event where it was set. There is no need to
retain the flag beyond that point as it isn't used again.
Add a new job attribute ORTE_JOB_FULLY_DESCRIBED to indicate that all the job information (including locations and binding) is included in the launch message. Thus, the backend daemons do not need to do any map computation for the job. Use this for the seq, rankfile, and mindist mappers until someone decides to update them.
Note that this will maintain functionality, but means that users of those three mappers will see large launch messages and less performant scaling than those using the other mappers.
Have the mindist module add procs to the job's proc array as it is a fully described module
Protect the hnp-not-in-allocation case
Per path suggested by Gilles - protect the HNP node when it gets added in the absence of any other allocation or hostfile
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
On unmanaged allocations, we need to update the total_slots_allocated once the daemons have been launched and "discovered" their topology
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Do not pass param file MCA params if the user has requested that no param files be read - required when trying to avoid launch time penalties from large numbers of processes reading default param files. The daemon picks them up and passes them along anyway, so it isn't clear what value we gain from having them all read the defaults
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
