Revamp the event notification integration to rely on the PMIx event chaining and remove the duplicate chaining in OPAL. This ensures we get system-level events that target non-default handlers.
Restore the hostname entries for MPI-level error messages, but provide an MCA param (orte_hostname_cutoff) to remove them for large clusters where the memory footprint is problematic. Set the default at 1000 nodes in the job (not the allocation).
Begin first cut at memory profiler
Some minor cleanups of memprobe
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Plug a minor memory leak. Tell the PMIx server not to create a dstore memory region for the daemon job as there is nobody to share it with.
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Protect users of hwloc membind functions
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Update PMIx to include NULL string protection
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Update to PMIx master to include key overwrite protection
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
There are only five places in the non-daemon code paths where opal_hwloc_topology is currently referenced:
* shared memory BTLs (sm, smcuda). I have added a code path to those components that uses the location string
instead of the topology itself, if available, thus avoiding instantiating the topology
* openib BTL. This uses the distance matrix. At present, I haven't developed a method
for replacing that reference. Thus, this component will instantiate the topology
* usnic BTL. Uses the distance matrix.
* treematch TOPO component. Does some complex tree-based algorithm, so it will instantiate
the topology
* ess base functions. If a process is direct launched and not bound at launch, this
code attempts to bind it. Thus, procs in this scenario will instantiate the
topology
Note that instantiating the topology on complex chips such as KNL can consume
megabytes of memory.
Fix pernode binding policy
Properly handle the unbound case
Correct pointer usage
Do not free static error messages!
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
MPI_T_pvar_get_index was returning an incorrect index. The index
was never set correctly while registering the performance variables.
Additionally fix a missing case in the mca_base_var_type_t to MPI
datatype conversion. This type is currently used for control variables
registered by mxm, fca and hcoll components.
Signed-off-by: Nysal Jan K.A <jnysal@in.ibm.com>
- Fix capitolization typos
- Make comment more correct / flow better
- Use AM_CPPFLAGS, not DEFAULT_INCLUDES
- Remove extra "hwloc/" from external hwloc.h specification
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
- simply #include "hwloc.h" to use the external hwloc header
- do use the external hwloc header instead of opal/mca/hwloc/hwloc.h
Thanks Orion Poplawski for the report
Fixesopen-mpi/ompi#2616
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
Fine tuning of flux component
Fix a few minor issues with the initial cut:
* Job id could be obtained from the PMI kvsname like SLURM,
but simpler to getenv (FLUX_JOB_ID)
* Flux pmi-1 doesn't define PMI_BOOL, PMI_TRUE, PMI_FALSE
* Flux pmi-1 maps the deprecated PMI_Get_kvs_domain_id() to
PMI_KVS_Get_my_name() internally, so just call that instead.
* Drop residual slurm references.
Add wrappers for PMI functions so that if HAVE_FLUX_PMI_LIBRARY
is not defined, the component can dlopen libpmi.so at location
specified by the FLUX_PMI_LIBRARY_PATH env variable, which adds
flexibility. If HAVE_FLUX_PMI_LIBRARY is defined, link with
libpmi.so at build time in the usual way.
Update configury for flux component
Update m4 so the configure options work as follows:
--with-flux-pmi
Build Flux PMI support (default: yes)
--with-flux-pmi-library
Link Flux PMI support with PMI library at build
time. Otherwise the library is opened at runtime at
location specified by FLUX_PMI_LIBRARY_PATH environment
variable. Use this option to enable Flux support when
building statically or without dlopen support (default: no)
If the latter option is provided, the library/header is located at
build time using the pkg-config module 'flux-pmi'. Otherwise there
is no library/header dependency.
Handle the case where ompi is configured with --disable-dlopen
or --enable-statkc. In those cases, don't build the component
unless --with-flux-pmi-library is provided.
It is fatal if the user explicitly requests --with-flux-pmi but
it cannot be built (e.g. due to --disable-dlopen).
Add a schizo/flux component
Update schizo/flux component
Eliminate slurm-specific usage cases.
Since the module is only loaded if FLUX_JOB_ID is set, there are
only two cases to handle:
1) App was launched indirectly through mpirun. This is not yet
supported with Flux, but hook remains in case this mode is supported
in the future.
2) App was launched directly by Flux, with Flux providing
CPU binding, if any.
Fix up white space in pmix/flux component
Drop non-blocking fence from pmix:flux component
The flux PMI-1 library is not thread safe, therefore
register a regular blocking fence callback instead of the
thread-shifting fencenb().
pmix/flux component avoids extra PMI_KVS_Gets
Keys stored into the base cache under the wildcard
rank are not intended to be part of the global key namespace.
These keys therefore should not trigger a PMI_KVS_Get() if they
are not found in the cache.
Minor pmix/flux component cleanup
pmix/flux: drop code for fetching unused pmix_id
pmix/flux: err_exit must return error
Problem: in flux_init(), although 'ret' (variable holding
err_exit return code) is initialized to OPAL_ERROR, the
variable is reused as a temporary result code, so if there are
some successes followed by a failure that doesn't set 'ret',
flux_init() could return success with PMI not initialized.
Ensure that a "goto err_exit" returns OPAL_ERROR if 'ret'
is not set to some other error code.
pmix/flux: don't mix OPAL_ and PMI_ return codes
Problem: flux_init() can return both PMI_ and OPAL_ return
codes. Although OPAL_SUCCESS and PMI_SUCCESS are both defined
as 0, other codes are not compatible.
Ensure that flux_init() consistently uses 'rc' for PMI_
return codes and 'ret' for OPAL_ return codes.
pmix/flux: factor out repeated code for cache put
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Newer x86 processors have a core invariant tsc. On these systems it is
safe to use the rtdtsc instruction as a monotonic timer. This commit
adds a new function to the opal timer code to check if the timer
backend is monotonic. On x86 it checks the appropriate bit and on
other architectures it parrots back the OPAL_TIMER_MONOTONIC value.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Update ORTE support for dynamic PMIx operations e.g., PMIx_Spawn
Update to track master
Ensure that --disable-pmix-dstore actually disables the dstore. Sync to a few debugger updates
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
It turns that there is an incompatibility between the Cray PMI
library and the default configuration for building Open MPI (master).
To work around this, we now disable use of aprun for direct launch
of Open MPI jobs except under specific conditions.
The problem is that there are now (on master) packages getting
initialized that do not work properly across a fork operation.
As part of a constructor in the Cray PMI library, a fork operation
is done to simplify use of shared memory between the
processes in a job on the same node. This ends up thoroughly
messing up the Open MPI initialization process in the case
that dlopen support is enabled. The initialization process gets
about half-way through when the PMIX framework is opened and
components are loaded, which triggers the Cray PMI constructor
and hence the fork operation.
There are two workarounds for this:
1) configure Open MPI for Cray XE/XC systems using aprun with the
--disable-dlopen option
2) set the PMI_NO_FORK environment variable in the shell in which
the aprun command is run.
Without taking these measures, a Open MPI job will just hang at
job startup in the first attempt to "thread-shift" the PMIx
fence_nb operation. Additional hangs occur at shutdown if this
problem is worked around, again due to the insertion of a fork
operation halfway through the Open MPI initialization procedure.
This commit detects if the conditions that bring out the hang
situation are present, and if so, prints out a message and
aborts the job launch.
Note on systems using slurm, the PMI_NO_FORK environment variable
is set as part of the srun job launch, hence this issue is avoided
on those systems.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
An file might have been destroyed by an other task between
readdir() and stat(), so simply ignore stat() failure.
That typically occurs when one task is removing the job_session_dir
and an other task is still removing its proc_session_dir.
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
opal_convertor_pack() might pack less bytes than requested,
so always set frag->segments[0].seg_len.
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
PR open-mpi/ompi#2432 introduced a regression where configure
and build with --disable-dlopn caused build failure owing
to unresolved alps lli symbols in the libopal-pal shared library.
This commit fixes this problem.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
Enhance the cray pmix component to set some OMPI internal
env. variables used to set some key/value pairs
on the MPI_INFO_ENV object. This allows more of the
ompi-tests ibm unit tests to pass when using aprun/srun
direct launch and Cray PMI.
Signed-off-by: Howard Pritchard <howardp@lanl.gov>
since pthreads are now mandatory, the MCA_BTL_TCP_SUPPORT_PROGRESS_THREAD
is always true and hence can be safely removed
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
Remove BTL_OPENIB_FAILOVER_ENABLED code in the openib btl source.
Remove the failover-specific files from the openib btl.
Update the openib/Makefile.am accordingly.
Remove the -enable-openib-failover config logic.
Signed-off-by: Steve Wise <swise@opengridcomputing.com>
This commit rewrites much of the btl/self component to fix a long
standing memory usage bug. Before this commit the prepare_src path
would always allocate a max send fragment (256kB). This caused the
rank to allocate 32 * 256k useless buffers from one send. This commit
makes the following changes:
- Add the MCA_BTL_FLAGS_GET flag by default. No reason not to set it.
- Reduce the eager limit, max send size, buffers per allocation, and
maximum buffer count per fragment size. These changes should have
no noticible affect on performance but should greatly reduce the
memory usage of the component.
- Implement the sendi function. This should reduce self send latency
somewhat.
- Rewrite prepare_src to never allocate a eager or max send fragment
for contiguous data.
- add_procs needs to return something in the peer array for the proc
self not just set the reachability bit. Now stores (void *) 1.
- Various cleanups. Removed and unused file.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The vader btl kept a per-peer registration cache to keep track of
attachments. This is not really a problem with small numbers of local
ranks but can be a problem with large SMP machines. To reduce the
footprint there is now one registration cache for all xpmem
attachments. This will probably increase the lookup time for large
transfers but is a worthwhile trade-off.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
the class system can be initialized/finalized as many times as we like,
so there is no more need to have opal_class_finalize() invoked in a destructor
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
libnl and libnl-3 are known to conflict with each other, so detect
and abort if these two libs are both used directly (e.g. Open MPI
uses libnl-3) or indirectly (e.g. libibverbs.so might depend on libnl)
Still not completely done as we need a better way of tracking the routed module being used down in the OOB - e.g., when a peer drops connection, we want to remove that route from all conduits that (a) use the OOB and (b) are routed, but we don't want to remove it from an OFI conduit.
OMPI header cannot be included in OPAL source code, hence removed it.
Fixes: (740b636db) btl/openib: Disqualify rdmacm CPC if
MPI_THREAD_MULTIPLE.
Signed-off-by: Potnuri Bharat Teja <bharat@chelsio.com>
if MPI_Init[_thread]/MPI_Finalize and MPI_T_init_thread/MPI_T_finalize
are balanced, opal_initialized is zero, and hence opal_cleanup destructor
never invokes opal_class_finalize.
if MPI_Init[_thread] nor MPI_T_init_thread have been called, classes is NULL,
so opal_class_finalize does nothing
- replace MAXHOSTNAMELEN with hardcoded 1024.
unlike Linux, Solaris #define MAXHOSTNAMELEN in <netdb.h>,
so use a hard coded value to keep the test simpl
- stdout cannot be assigned on Solaris, so use freopen instead
(back-ported from upstream commit pmix/master@a63f6e53f4)
this is a convenience macro similar to the PMIX_LIST_FOREACH macro,
that can be used to iterate on all the key/value pairs of a pmix_hash_table_t
(back-ported from upstream commit pmix/master@349971c68c)
We only support running with libfabric v1.3 or greater. So it's safe
to remove the legacy/adaptive cq_readerr() behavior.
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
There are critical usnic libfabric AV insert bugs before v1.3, so
don't allow any version prior to v1.3 at run time (still allow
*compiling* with earlier versions, though, since the ABI guarantees
allow us to compile with an earlier libfabric and run with a later
libfabric).
Switch to using fi_version() to check the version (instead of calling
fi_getinfo()) as a potentially lighter-weight / simpler solution.
This allows us to only call fi_getinfo() once.
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
The (effective) "+42" computation was, in fact, the incorrect answer
in this case (gasp!).
We should just take the max_msg_size from the command (which came from
the libfabric endpoint max_msg_size attribute in the client) and
subtract off the max header size: 68 (which is explained in the
comment). This will result in a "large" message size which is likely
slightly smaller than the MTU, but still right up near the MTU, and
therefore good enough.
Note: the old computation (i.e., -(68-42)) worked fine when we asked
for Libfabric API v1.1 because the usnic provider would return a
max_msg_size that was already less than the MTU due to FI_PREFIX
behavior shenanigans. Once we started asking for Libfabric API v1.4,
the usnic Libfabric provider started returning (MTU + prefix_size),
and the -(68-42) computation started giving a value that was over the
MTU. This caused sendto() on the connectivity checker UDP socket
to fail.
This commit also removes an old/misleading comment.
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
* Add a configure time option to rename libmpi(_FOO).*
- `--with-libmpi-name=STRING`
* This commit only impacts the installed libraries.
Internal, temporary libraries have not been renamed to limit the
scope of the patch to only what is needed.
For example:
```shell
shell$ ./configure --with-libmpi-name=wookie
...
shell$ find . -name "libmpi*"
shell$ find . -name "libwookie*"
./lib/libwookie.so.0.0.0
./lib/libwookie.so.0
./lib/libwookie.so
./lib/libwookie.la
./lib/libwookie_mpifh.so.0.0.0
./lib/libwookie_mpifh.so.0
./lib/libwookie_mpifh.so
./lib/libwookie_mpifh.la
./lib/libwookie_usempi.so.0.0.0
./lib/libwookie_usempi.so.0
./lib/libwookie_usempi.so
./lib/libwookie_usempi.la
shell$
```
theses three pmix components use the same class name,
declare it as static so Open MPI can be built with --disable-dlopen
Thanks Limin Gu for the report
The rdmacm CPC in the openib BTL is not thread safe. The rdmacm CPC
should disqualify itself (instead of failing in random ways) if
MPI_THREAD_MULTIPLE is the thread level.
Signed-off-by: Potnuri Bharat Teja <bharat@chelsio.com>
This patch is based on the "RFC: Reenabling the TCP BTL over local
interfaces (when specifically requested)". It removes the hardcoded
exception for the local devices that has been enforced by the
TCP BTL. Instead, we exclude the local interface only via the
exclude MCA (both IPv4 and IPv6 local addresses are already in the
default if_exclude), which is also the behavior currently described in
our README file.
pmix_progress_thread_finalize() invokes libevent event_base_free,
so all libevent stuff cannot be used after.
Hence, pmix_client_globals.myserver must be PMIX_DESTRUCT'ed
before invoking pmix_progress_thread_finalize()
Prevent a race condition between a thread checking count and then
going in cond_wait, and another thread setting the count to 0 and
signaling the condition.
Thanks to Pascal Deveze for catching up the bug and for
the initial patch.
pmix1_value_unload() was added a "key" argument which is unused,
and pmix1_value_unload() was sometimes invoked with two arguments instead of three.
since the "key" argument is unused, simply remove it from the
subroutine prototype and calls.
The add_64, sub_64, and cmpset_64 atomics used "+m" (*addr) to
indicate the asm also writes the memory location. This is better than
using a memory clobber. PGI 16.9 introduced a bug that causes a
compiler failure on the "+m" constraint (input/output). It seems to
work with "=m" (output) which matches the 32-bit atomics.
Fixesopen-mpi/ompi#2086
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
the LT_* macros do overwrite the enable_dlopen variable,
so it must be tested and saved before invoking LT_INIT.
delay the invokation of the LT_* macros and use the
PMIX_ENABLE_DLOPEN_SUPPORT variable to figure out whether
--disable-dlopen was invoked
This commit fixes an abort during finalize because pending events were
removed from the list twice.
References #2030
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
WAIT_SYNC_INIT(sync,0); WAIT_SYNC_RELEASE(sync);
hanged because sync->signaled was initialised to true, and
there is no reason to invoke WAIT_SYNC_SIGNALED(sync) before
WAIT_SYNC_RELEASE(sync)
this commit initializes sync->signaled to true unless the count is zero.
Thanks George for the review and guidance.
We commonly see messages on the users list where a peer has hung up
because it has crashed. Instead of having just a BTL_ERROR message,
make this a real opal_show_help() message that tells the user that the
peer unexpectedly hung up, and they should look into *why* that peer
hung up.
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
This commit adds selective use of a compiler-specific pragma to
silence the numerous warnings the Sun/Oracle/Studio compilers emit for
the GNU-style inline asm used in atomic.h.
Thanks Paul Hargrove for the initial patch and the guidance.
This commit contains the following changes:
- There is a bug in the PGI 16.x betas for ppc64 that causes them to
emit the incorrect instruction for loading 64-bit operands. If not
cast to void * the operands are loaded with lwz (load word and
zero) instead of ld. This does not affect optimized mode. The work
around is to cast to void * and was implemented similar to a
work-around for a xlc bug.
- Actually implement 64-bit add/sub. These functions were missing and
fell back to the less efficient compare-and-swap implementations.
Thanks to @PHHargrove for helping to track this down. With this update
the GCC inline assembly works as expected with pgi and ppc64.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
It looks like one help message was accidentally pasted in the middle
of another. Disentangle the two messages from each other, and
slightly tweak the one message to say that the job may also crash (in
addition to hanging).
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
This commit prevents the connection code from trying to connect an
endpoint if the directed datagram has been posted but not received.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The xlc compiler seems to behave in a different way that gcc when it
comes the inline asm. There were two problems with the code with xlc:
- The TOC read in mca_patcher_base_patch_hook used the syntax
register unsigned long toc asm("r2") to read $r2 (the TOC
pointer). With gcc this seems to behave as expected but with xlc
the result in toc is not the same as $r2. I updated the code to use
asm volatile ("std 2, %0" : "=m" (toc)) to load the TOC pointer.
- The OPAL_PATCHER_BEGIN macro is meant to be the first thing in a
hook. On PPC64 it loads the correct TOC pointer (thanks to
mca_patcher_base_patch_hook) and saves the old one. The
OPAL_PATCHER_END macro restores the TOC pointer. Because we *need*
the TOC to be correct before it is accessed in the hook the
OPAL_PATCHER_BEGIN macro MUST come first. We did this and all was
well with gcc. With xlc on the other hand there was a TOC access
before the assembly inserted by OPAL_PATCHER_BEGIN. To fix this
quickly I broke each hook into a pair of function with the
OPAL_PATCHER_* macros on the top level functions. This works around
the issue but is not a clean way to fix this. In the future we
should 1) either update overwrite to not need this, or 2) figure
out why xlc is not inserting the asm before the first TOC read.
This fixesopen-mpi/ompi#1854
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
With libfabric v1.4, the usnic provider changed the values of its
fabric and domain name strings (compared to libfabric <v1.4). Update
the Open MPI usNIC BTL to handle both pre-v1.4 and v1.4 fabric/domain
names.
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
and fail with a user friendly message if no method is available:
"sec: native cannot validate_cred on this system"
(back-ported from upstream pmix/master@c474a1fc60)
