Turns out there are some cases where the Cray
wlm_detect_get_active may return NULL, in which
case fallback to wlm_detect_get_default method
is suggested. Make use of the fallback to
avoid segfaults under some circumstances in the
ALPS plm selection method.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
* qos framework is moving to the scon layer and is no longer required in ORTE
* remove the rml/ftrm component as we now have multiple active components, and so the wrapper needs to be rethought
* no need for separating the "base" from "API" module definition. The two are identical
* move the "stub" functions into their own file for cleanliness
* general cleanup to meet coding standards
* cleanup some logic in the stubs
* provide a more reliable way of determining that a process is a singleton by leveraging the schizo framework. Add new components for slurm, alps, and orte to detect when we are in a managed environment, and if we have been launched by mpirun or a native launcher. Set the correct envars to control ess and pmix selection in each case.
* change the relative priority of the pmix120 and pmix112 components to make pmix120 the default
* fix singleton comm-spawn by correctly setting the num_apps field of the orte_job_t created by the daemon - this fixes a segfault in register_nspace on newly created daemons
* ensure orterun doesn't propagate any ess or pmix directives in its environment
* Cleanup a few valgrind issues and memory leaks
* Fix a race condition that prevented the client from completing notification registrations (missing thread shift)
* Ensure the shizo/alps component detects launch by mpirun
There is a bug in MPMD detection that disables totalview if a : is
found anywhere on the command line. This includes inside an argument
option or MCA variable value. This commit changes the check to look
for the string " : " instead of the character : which should eliminate
the issue in most cases.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
If ibv_fork_init() has been invoked the pages are marked MADV_DONTFORK.
If we only partially use a page, any data allocated on the remainder of
the page will be inaccessible to the child process.
Fixesopen-mpi/ompi#1363
* Clean up the DVM so it continues to run even when applications error out and we would ordinarily abort the daemons.
* Create a new errmgr component for the DVM to handle the differences.
* Cleanup the DVM state component.
* Add ORTE bindings directory and brief README
* Pass a local tool index around to match jobs.
* Pass the jobid on job completion.
* Fix initialization logic.
* Add framework for python wrapper.
* Fix terminate-with-non-zero-exit behavior so it properly terminates only the indicated procs, notifies orte-submit, and orte-dvm continues executing.
* Add some missing options to orte-dvm
* Fix a bug in -host processing that caused us to ignore the #slots designator. Add a new attribute to indicate "do not expand the DVM" when submitting job spawn requests.
* It actually makes no sense that we treat the termination of all children differently than terminating the children of a specific job - it only creates confusion over the difference in behavior. So terminate children the same way regardless.
Extend the cmd_line utility to easily allow layering of command line definitions
Catch up with ORTE interface change and make build more generic.
Disable "fixed dvm" logic for now.
Add another cmd_line function to merge a table of cmd line options with another one, reporting as errors any duplicate entries. Use this to allow orterun to reuse the orted_submit code
Fix the "fixed_dvm" logic by ensuring we reset num_new_daemons to zero. Also ensure that the nidmap is sent with the first job so the downstream daemons get the node info. Remove a duplicate cmd line entry in orterun.
Revise the DVM startup procedure to pass the nidmap only once, at the startup of the DVM. This reduces the overhead on each job launch and ensures that the nidmap doesn't get overwritten.
Add new commands to get_orted_comm_cmd_str().
Move ORTE command line options to orte_globals.[ch].
Catch up with extra orte_submit_init parameter.
Add example code.
Add documentation.
Bump version.
The nidmap and routing data must be updated prior to propagating the xcast or else the xcast will fail.
Fix the return code so it is something more expected when an error occurs. Ensure we get an error returned to us when we fail to launch for some reason. In this case, we will always get a launch_cb as we did indeed attempt to spawn it. The error code will be returned in the complete_cb.
Fix the return code from orte_submit_job - it was returning the tracker index instead of "success". Take advantage of ORTE's pretty-print capabilities to provide a nice error output explaining why we failed to launch. Ensure we always get a launch_cb when we fail to launch, but no complete_cb as the job never launched.
Extend the error reporting capability to job completion as well.
Add index parameter to orte_submit_job().
Add orte_job_cancel and implement ORTE_DAEMON_TERMINATE_JOB_CMD.
Factor out dvm termination.
Parse the terminate option at tool level.
Add error string for ORTE_ERR_JOB_CANCELLED.
Add some safeguards.
Cleanup and/of comments.
Enable the return.
Properly ORTE_DECLSPEC orte_submit_halt.
Add orte_submit_halt and orte_submit_cancel to interface.
Use the plm interface to terminate the job
Cleanup the configury so we properly check for Singularity under the various typical use-cases
Bring the Singularity support online. We have to turn "off" the sm BTL as it segfaults from inside the container - root cause remains unclear. Also turned "off" the various OPAL shmem components in case they are involved and someone else tries to use them. Happily, the vader BTL works just fine!
converting an opal_process_name_t means the loss of one bit,
it was decided to restrict the local job id to 15 bits, so the
useful information of an opal_process_name_t can fit in 63 bits.
This commit fixes multiple issues in the bruck's and recursive
doubling grpcomm algorithms. The following changes are included:
- Use the existing bitmap implementation instead of implementing a
new one. There were bugs in the implementation that caused an
overrun of the bitmap array.
- Clean up the algorithms to eliminate errors.
- Send as little extra data as possible in the bruck's
algorithm.
The changes were testest with various numbers of ortes varying from 1
to 4096.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Update the configure logic for the new pmix120 component
ckpt
Get the pmix120 component to work - still not really registering or handling notifications, but infrastructure now operates
Cleanup some of the symbol scopes, and provide a more comprehensive rename.h file. Will pretty it up later - let's see how this works
Cleanup the rename files to use the pretty macros
Turns out that the way the SLURM plm works
is not compatible with the way MPI processes
on Cray XC obtain RDMA credentials to use
the high speed network. Unlike with ALPS,
the mpirun process is on the first compute
node in the job. With the current PLM launch
system, mpirun (HNP daemon) launches the MPI
ranks on that node rather than relying on
srun.
This will probably require a significant amount
of effort to rework to support Native SLURM
on Cray XC's. As a short term alternative,
have the alps plm (which gets selected by default
again on Cray systems regardless of the launch system)
check whether or not srun or alps is being used on the
system. If alps is not being used, print a helpful
message for the user and abort the job launch.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
Move .so version numbers to their appropriate project in the top-level
VERSION file. Also add the project name to all .so version number
names. Remove no-longer-used .so names.
NOTE: Building with external pmix *requires* that you also build with external libevent and hwloc libraries. Detect this at configure and error out with large message if this requirement is violated.
Closes#1204 (replaces it)
Fixes#1064
Due to user confusion, update the show-help messages displayed when
processor and/or memory binding fails. Thanks to Dave Love
(@loveshack) for the initial suggestion.
Fixesopen-mpi/ompi#1087
Rename the pmix1xx component to pmix111 so it reflects the actual release it includes
Resolve the problem of PMIx being passed a bogus --with-platform argument when configuring the PMIx tarball code. There is no reason we should be passing --with-platform arguments to any internal subdirectory, so just leave that out when constructing the opal_subdir_args variable.
Update the PMIx code and continue attempting to debug direct modex
Fix a problem in the ORTE PMIx server - there was an early intent to optimize the direct modex by fetching data for all procs from the target job on the remote node, instead of fetching the data one proc at a time. However, this was never completely implemented, and so we would hang if we had multiple overlapping requests for data from more than one proc on the node.
Update PMIx to v1.1.2
Cray has added plugins to slurm to support
the Cray programming env (alpslli, cray pmi, etc).
Some of the workarounds needed with plm/alps
to avoid issues with Cray PMI getting mixed up
with orte launch system are also required in
a cray native slurm environment.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
to continue current default behavior.
Also add an MCA param pmix_base_collect_data to direct that the blocking fence shall return all data to each process. Obviously, this param has no effect if async_
modex is used.
The mca_base_select function uses returned priorities to select the
best component/module. This priority may be of use to the caller so
pass that information back in an optional argument. If the priority is
not needed pass NULL.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This required modifying the mca_component_select function to actually check the return code on a component query - it was blissfully ignoring it.
Also do a little cleanup to avoid bombarding the user with multiple error messages.
Thanks to Patrick Begou for reporting the problem
Bring Slurm PMI-1 component online
Bring the s2 component online
Little cleanup - let the various PMIx modules set the process name during init, and then just raise it up to the ORTE level. Required as the different PMI environments all pass the jobid in different ways.
Bring the OMPI pubsub/pmi component online
Get comm_spawn working again
Ensure we always provide a cpuset, even if it is NULL
pmix/cray: adjust cray pmix component for pmix
Make changes so cray pmix can work within the integrated
ompi/pmix framework.
Bring singletons back online. Implement the comm_spawn operation using pmix - not tested yet
Cleanup comm_spawn - procs now starting, error in connect_accept
Complete integration
* don't pass --tree-spawn to the orted cmd line. If someone doesn't want tree-spawn, it shows up as an MCA param anyway
* ensure state/orted component disqualifies itself from CM operations
* clarify the DVM proc_type definitions
* ensure we stop littering the tmp dir with session directories
In order to address issue #741, the orted's now are
always launched with the Cray PMI environment variables
PMI_NO_FORK
PMI_NO_PREINITIALIZE
set to disable running of the library's ctor.
So there's no longer a need to set these for the
application(s) being launched by the orted's.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
Turns out that when one builds Open MPI with --disable-dlopen
for Cray, a whole bunch of cray specific libraries get linked
in to the orted executable. One of these is Cray PMI. The
Cray PMI has a ctor which, if run, causes job launches using
mpirun to fail. This commit suppresses the running of the
ctor and thus prevents failure to launch.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
This commit does two things. It removes checks for C99 required
headers (stdlib.h, string.h, signal.h, etc). Additionally it removes
definitions for required C99 types (intptr_t, int64_t, int32_t, etc).
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
Changing the client to leave its socket as blocking during the connect doesn't solve the problem by itself - you also have to introduce a sleep delay once the backlog is hit to avoid simply machine-gunning your way thru retries. This gets somewhat difficult to adjust as you don't want to unnecessarily prolong startup time.
We've solved this before by adding a listening thread that simply reaps accepts and shoves them into the event library for subsequent processing. This would resolve the problem, but meant yet another daemon-level thread. So I centralized the listening thread support and let multiple elements register listeners on it. Thus, each daemon now has a single listening thread that reaps accepts from multiple sources - for now, the orte/pmix server and the oob/usock support are using it. I'll add in the oob/tcp component later.
This still didn't fully resolve the SMP problem, especially on coprocessor cards (e.g., KNC). Removing the shared memory dstore support helped further improve the behavior - it looks like there is some kind of memory paging issue there that needs further understanding. Given that the shared memory support was about to be lost when I bring over the PMIx integration (until it is restored in that library), it seemed like a reasonable thing to just remove it at this point.
CID 71993 Logically dead code (DEADCODE)
As indicated by coverity proc can not be NULL at any point after the
continue. Removed dead code.
CID 1269682 Unchecked return value (CHECKED_RETURN)
Check the return code of orte_get_attribute. I assume we still need to
check for a NULL proc in case the aborted proc attribute is set to
NULL. This might be better as an assert ().
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Have only a single level of "if" conditionals. Also, slightly change
the logic such that we only die/break out of the loop if we get EMFILE
-- all other errors are ok to go on to the next fd.
Finally, use a real show_help() message to warn when other errors occur.
The timeout is frequency at which to send keepalive pings; the retry
interval is how often to send successive pings once a keepalive has
not replied.
Also update comments and MCA param help strings.
60 seconds -- squashme
It is perfectly ok to be on a system without UD devices.
Also, make some of the error messages better -- so that the user has a
clue about where the error messages are coming from, and what they
should do.
There was a redundant computation of the vpid
for orted's happening in ess/alps rte_init
method. Keep the more efficient alps based
method.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
A few uninitialized common symbols are remaining (generated by flex) :
* orte/mca/rmaps/rank_file/rmaps_rank_file_lex.c: orte_rmaps_rank_file_leng
* orte/mca/rmaps/rank_file/rmaps_rank_file_lex.c: orte_rmaps_rank_file_text
* orte/util/hostfile/hostfile_lex.c: orte_util_hostfile_leng
* orte/util/hostfile/hostfile_lex.c: orte_util_hostfile_text
Move the call to opal_common_verbs_fork_test() to up before the call
to ibv_get_device_list() (just curious -- why not use
opal_ibv_get_device_list()?). This ensures that the call to
ibv_fork_init() is before *all* other ibv_* calls.
hwloc output can get fairly long, especially on machines with lots of
cores and/or hyperthreads. So put the Locale and Binding output on
separate lines.
The oob/ud configure was not honoring the case
if the ompi is configured with --with-verbs=no.
This fixes that problems.
Fixes#522
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
This commit fixes several vagrind errors. Included:
- installdirs did not correctly reinitialize all pointers to NULL
at close. This causes valgrind errors on a subsequent call to
opal_init_tool.
- several opal strings were leaked by opal_deregister_params which
was setting them to NULL instead of letting them be freed by the
MCA variable system.
- move opal_net_init to AFTER the variable system is initialized and
opal's MCA variables have been registered. opal_net_init uses a
variable registered by opal_register_params!
- do not leak ompi_mpi_main_thread when it is allocated by
MPI_T_init_thread.
- do not overwrite ompi_mpi_main_thread if it is already set (by
MPI_T_init_thread).
- mca_base_var: read_files was overwritting mca_base_var_file_list
even if it was non-NULL.
- mca_base_var: set all file global variables to initial states on
finalize.
- btl/vader: decrement enumerator reference count to ensure that it
is freed.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
putenv requires that any string put into the environment is not
changed or freed. That is not the case with constant strings as they
will go away when dlclose is called on the component. Instead, just
use opal_setenv which does not have this restriction.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit fixes the following bugs:
- opal_output_finalize did not properly set internal state. This
caused problems when calling the sequence opal_output_init (),
opal_output_finalize (), opal_output_init ().
- opal_info support called mca_base_open () but never called the
matching mca_base_close (). mca_base_open () and mca_base_close ()
have been updated to use a open count instead of an open flag to
allow mca_base_open to be called through multiple paths (as may be
the case when MPI_T is in use).
- orte_info support did not register opal variables. This can cause
orte-info to not return opal variables.
- opal_info, orte_info, and ompi_info support have been updated to
use a register count.
- When opening the dl framework the reference count was added to
ensure the framework stuck around. The framework being closed
prematurely was a bug in the MCA base that has since been
corrected. The increment (and associated decrement) have been
removed.
- dl/dlopen did not set the value of
mca_dl_dlopen_component.filename_suffixes_mca_storage on each call
to register. Instead the value was set in the component
structure. This caused the value to be lost when re-loading the
component. Fixed by setting the default value in register.
- Reset shmem framework state on close to avoid returning a stale
component after reloading opal/shmem.
- MCA base parameters were not properly deregistered when the MCA
base was closed.
This commit may fix#374.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit adds support for project_framework_component_* parameter
matching. This is the first step in allowing the same framework name
in multiple projects. This change also bumps the MCA component version
to 2.1.0.
All master frameworks have been updated to use the new component
versioning macro. An mca.h has been added to each project to add a
project specific versioning macro of the form
PROJECT_MCA_VERSION_2_1_0.
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
Don't filter the topology by cpuset if you are mpirun until you know that no other compute nodes are involved. This deals with the corner case where mpirun is executing on a node of different topology from the compute nodes.
Simplify - don't mandate that all cpus in the given cpuset be present on every node. We can then run everything thru the filter as before, which ensures that any procs run on mpirun are also contained within the specified cpuset.
Correctly count the number of available PUs under each object when given a cpuset
Fix the default binding settings, and correctly count PUs when no cpuset is given
Ensure the binding policy gets set in all cases
Have to have alps rpms installed on a system
for alps component to build, even if separated
by a level of indirection.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
Implement an almost-do-nothing alps oob component.
When using aprun to launch a job on Cray system,
there is no reason to need an oob system, since ompi
relies on Cray PMI for oob communication.
Fixes#484
Enabling the FT code breaks compilation (again). This series
tries to fix the compiler errors. This is again only fixing
the compiler errors without any warranty that the result
might actually support FT again.
With the changes introduced in the previous patches in this series
some goto constructs for cleanup are no longer necessary and removed.
