- Fix bug identified by users: --prefix may also apply on the local
node; we need to prefix the PATH and LD_LIBRARY_PATH environment
variables before invoking execve()
This commit was SVN r7541.
some orted's to stall on locks in the MPI Dynamics cases. Since it
is not essentual that we call these functions, they can so away.
Unlock the peer lock when aborting. This causes a potential deadlock
in do_waitall [see comment in code]. This was causing orteds to
deadlock at times when the seed had terminated. With proper interleaving
and timing the orted was deadlocking. This seems to have fixed this in
my stress testing with MPI 2 Dynamics.
This commit was SVN r7539.
The NS replica should give out tags that are over ORTE_RML_TAG_DYNAMIC
or it will overlap with other outstanding tags. This overlap was killing
MPI_Comm_spawn when a program tried to use it multiple times (> 3).
With this fix MPI_Comm_spawn is behaving properly.
A program can call it many times in a row with out problem.
NOTE: Not tested for multi-threaded build yet
(A long time debugging for a one liner... :/)
This commit was SVN r7529.
Have the ras_base_schedule_policy MCA parameter working once again. before it
would only do slot based allocation, even if the MCA parameter was set properly.
Currently you can specify to orterun a node allocation by either:
-mca ras_base_schedule_policy node
-bynode
and slot allocation (which is the default) by:
-mca ras_base_schedule_policy slot
-byslot
This commit was SVN r7513.
with the mutex locked and as this function will call oob_send which will call the lookup again
... we will deadlock as the mutex is already lock. The solution is to release the mutex before
going into the subscription. Then of course the logic to remote the item when something went
wrong with the subscrition is a little bit more complex.
This commit was SVN r7429.
add an event, it can call the spawn function directly. This will avoid it standing on the condition who
will never get released.
This commit was SVN r7428.
This fixes one of the race conditions in orterun is sent a kill signal.
Before it would sometimes spin in the OOB waiting for a message to complete
to a peer that was no longer around. Stalling at this level prevented orterun
from noticing that it had received a kill signal.
This commit was SVN r7408.
automagically don't build on platforms without such things
* Fix for mistaken use of cache variable in assembly setup
* one more cached test hits the books
This commit was SVN r7404.
LIBADD instead of appending to the existing one.
Also removed some more Makefile.options whitespace, and I think emacs
removed some tabs (i.e., replaced them with whitespace).
This commit was SVN r7399.
Makefile.options
- Sample in each of the three projects of how to link againt the
relevant libraries so that when components are loaded into a parent
process' space, we don't rely on the libopal/liborte/libmpi symbols
being in the parent's public symbol namespace -- instead,
dynamically link to the relevant libraries, allowing the dynamic
linker to pull those libraries in at run-time, if needed
This commit was SVN r7397.
However we do want to do a bit of cleanup on the node before we exit,
specificly clean out the session directory. I also had a couple of the
subsystems that don't depend upon peers (which is key) clean up as well.
Pedantic formatting issue in oob_tcp.h
This commit was SVN r7387.
caller to specify a subset of the state variables that it can can subscribe to.
This is specified with one of three special flags defined in rmgr/rmgr_types.h
This is useful when we only care about a subset of the state changes, such as
in orted which only needs to know when a job has terminated or aborted.
This commit was SVN r7356.
waiting instead for the SOH to indicate that the jobid has terminated.
In a scheduled environment, if your program has a section of MPI code
followed by a section of computation that some processes execute while
other proceses terminate normally. This patch keeps the scheduler from
terminating all of the processes and the allocation if all of the processes
on an allocated node exit well before other processes on other nodes.
This commit was SVN r7333.
The following formats are parsed:
user@IPv4
user@fqdn
IPv4 or fqdn [username|user-name|user_name]=user
- Try a better error-detection when parsing (recognize wrong
IPs, fqdns...)
This commit was SVN r7288.
that multiple processes don't overwrite each other. Change that
default in orte_init_stage1() to just "output-" (because the file will
be in a process-unique directory at that point; the pid is no longer
necessary).
This commit was SVN r7256.
opal_output_set_output_file_info(). This allows getting and setting
the default directory where output stream files will be opened (for
all *new* streams). Before this function is not invoked, the default
location is $TMPDIR or $HOME (if $TMPDIR is not defined).
Added a call into orte_init_stage1() to call this function
immediately after the session directory is created and set the default
location of stream files to be the process' session directory.
This commit was SVN r7254.
AM_INIT_AUTOMAKE, instead of the deprecated version.
* Work around dumbness in modern AC_INIT that requires the version
number to be set at autoconf time (instead of at configure time, as
it was before). Set the version number, minus the subversion r number,
at autoconf time. Override the internal variables to include the r
number (if needed) at configure time. Basically, the right thing
should always happen. The only place it might not is the version
reported as part of configure --help will not have an r number.
* Since AM_INIT_AUTOMAKE taks a list of options, no need to specify
them in all the Makefile.am files.
* Addes support for subdir-objects, meaning that object files are put
in the directory containing source files, even if the Makefile.am is
in another directory. This should start making it feasible to
reduce the number of Makefile.am files we have in the tree, which
will greatly reduce the time to run autogen and configure.
This commit was SVN r7211.
This allows the user to specify certain options to srun when an application
is launched with this PLS.
A useful example is the need to set the time to wait from when the first
process completes and when slurm kills remaining processes:
pls_slurm_args=--wait=1200
This commit was SVN r7206.
app_context:
mpirun -np 2 -prefix /path/to/ompi/on/machineA ./exec1 : \
-np 2 -prefix /path/to/ompi/on/machineB ./exec2
- Allow with -mca pls_rsh_assume_same_shell 0, the checking for the
SHELL-variable on the actual node (currently 1st node).
Sets the prefix, PATH and LD_LIBRARY_PATH for bash/ksh and
csh/tcsh.
This commit was SVN r7195.
CTRL-C'd.
We were calling orte_finalize recursively which caused a segv when it tried to
use a freed framework (orte_rmgr in this case).
I added a status flag to orte_universe_info to indicate where we are in the code.
This was needed to determine if we should call orte_abort or not when shutting
down in the tcp oob.
This commit was SVN r7160.
1. Valgrind is good for something - chasing down memory leaks in registry led me to re-visit the dictionary functions and discover that I wasn't keeping track of the number of dictionary entries on each segment! Resulted in wasted time searching blank entries as well as leaked memory. This has now been fixed.
2. Fixed the orte_bitmap test. The init function for that class has been eliminated and the constructor adjusted to provide that functionality.
This commit was SVN r7136.
1. user does NOT specify the universe name. For the default universe case, if we detect an existing default universe and cannot connect to it, we quietly create an alternative default name by adding the pid to the orte_default_universe name and move on - we no longer provide a warning message for this case.
2. user specified a universe name. If we detect an existing universe of that name and cannot connect to it, we consider this an error condition and abort.
This commit was SVN r7131.
1. Added OMPI_PROC_ARCH as a defined registry key and added the code so that the architecture info gets properly transmitted across all processes using the startup message.
2. Added an OMPI_MODEX_KEY definition and removed the hard-coded "modex" key from pml_modex_exchange
This commit was SVN r7129.
add a -I to find the included ltdl.h (vs. a system-installed ltdl.h)
- Clean up kruft in a bunch of Makefile.am's to remove now-unnecessary
AM_CPPFLAGS settings to get static-components.h for each framework
- Move the component_repository API functions out of opal/mca/base/base.h
and into opal/mca/base/mca_base_component_repository.h in order to
decrease unnecessary dependencies (e.g., before this, almost
everything in the tree depended on ltdl.h, which is unnecessary --
only a small number of files really need ltdl.h)
This commit was SVN r7127.
include any optimization flags
- Use these flags to always compile ompi/debuggers/* and orterun so
that parallel debuggers (such as Totalview) can always see the
debugging symbols (see comments in ompi/debuggers/Makefile.am and
orte/tools/orterun/Makefile.am)
- Remove some obsolete LAM-named variables from configure.ac
This commit was SVN r7125.
Here's the huge registry check-in you've all been waiting for with baited breath. The revised version sends a single message to all processes at the various stage gates, thus making the startup much more scalable. I could provide you with all the tawdry details, but won't for now - you are welcome to ask, though, and I'll merrily bore your ears to tears.
In addition, the commit contains the following:
1. set the ignore properties on ompi/debuggers and orte/mca/pls/poe
2. Added simplified subscribe and put functions to the registry's API. I have also converted all of the ompi functions that registered subscriptions to the new API, and caught their associated put's as well.
In a follow-on commit, I'll be adding support for George's hetero arch registry subscription (wanted to get this one in first).
This commit was SVN r7118.
it to be an exit.
* Put the srun process (or what is about to become the srun process) in
it's own process group so that group-wide signals (such as the
SIGINT sent by hitting cntl-c in a shell) are not sent to the srun
process.
This commit was SVN r7068.
orte_init_stage1(), since not all ORTE processes call orte_init().
* Expad opal_error test case to make sure ORTE error codes print
properly
* Make project error codes start at easy values (OPAL is -1 to -100,
ORTE is -101 to -200, OMPI is less than -201) to make it easier
to figure out what an error code as an integer means. Also has
the nice property of not changing the values of error codes ever
time a new error code is added.
This commit was SVN r7061.
tree.
- fix up #include's throughout the tree (yay contrib/search_replace.pl!)
- remove a few extraneous #include's
- remove orte_sys_info*() from opal_init()/opal_finalize() (it's
already in orte_init_stage1() and orte_system_finalize())
- remove dependencies in opal on orte_system_info -- util/os_path.c
and util/os_create_dirpath.c (they only used path_sep, anyway --
easily changed to #defines)
This commit was SVN r7059.
session directory cleanup (among other things)
- When we get an abnormal exit in orterun (i.e., timeout expires and
we haven't gotten termination notices from all processes), print a
better message an exit in a better way (which includes session
directory cleanup)
- Fix tm and poe pls's to not exit() but rather propagate the error up
the stack (where relevant)
This commit was SVN r7058.
- Change orte_base_infrastructre to orte_infrastructre to conform with
ompi_info's needs
- Move MCA Param registration in ORTE to a centralized function that is
called first in orte_init_stage1
- Set the infrastructre flag as an argument to orte_init
- Adjust initalization functions to properly pass down the infrastructre
flag.
This commit was SVN r7053.
Also check to see if infrastructre flag was previously set before assuming it
to be false. This was causing orterun to operate incorrectly in the presence
of a persistant daemon.
This commit was SVN r7039.
NOTE: These have NOT been added to the Makefile.am in the repository. Please do NOT add them at this time - I will do so later.
This commit was SVN r6979.
OPAL_ERROR, same for all the other error codes. Also, make sure that there
are never conflicts between OPAL anr ORTE error codes (for example).
Finally, provide opal_perror(), opal_strerror(), and opal_strerror_r() to
give stringified error messages for the different error codes
This commit was SVN r6969.
multi-client issues the old version had. Also, ignore the NULL iof
component, since we shouldn't use it when using the proxy orteds
This commit was SVN r6939.
against the total number of processors. If not oversubscribing, emit
the MCA environment variable mpi_paffinity_processor with the
processor number to bind the process to. This parameter is picked up
during MPI_Init (i.e., ompi_mpi_init()) and used to bind the process,
but currently iif the MCA param mpi_paffinity_alone is set to a
nonzero value (i.e., the user asks for it).
This commit was SVN r6906.
- change the framework opens to [mostly] use the new MCA param API
- properly pass in framework debug output streams to the
mca_base_component_open() function
This commit was SVN r6888.
* Add base to memory framework so that we can do something sane with
ompi_info
* Updated ompi_info to print components for memory framework and
show whether we have memory hooks active or not.
This commit was SVN r6861.
ns_replica.c
- Removed the error logging since I use this function in orte_init_stage1 to
check if we have created a cellid yet or not.
ras_types.h & rase_base_node.h
- This was an empty file. moved the orte_ras_node_t from base/ras_base_node.h
to this file.
- Changed the name of orte_ras_base_node_t to orte_ras_node_t to match the
naming mechanisms in place.
ras.h
- Exposed 2 functions:
- node_insert:
This takes a list of orte_ras_base_node_t's and places them in the Node
Segment of the GPR. This is to be used in orte_init_stage1 for singleton
processes, and the hostfile parsing (see rds_hostfile.c). This just puts
in the appropriate API interface to keep from calling the
orte_ras_base_node_insert function directly.
- node_query:
This is used in hostfile parsing. This just puts in the appropriate API
interface to keep from calling the orte_ras_base_node_query function
directly.
- Touched all of the implemented components to add reference to these new
function pointers
ras_base_select.c & ras_base_open.c
- Add and set the global module reference
rds.h
- Exposed 1 function:
- store_resource:
This stores a list of rds_cell_desc_t's to the Resource Segment.
This is used in conjunction with the orte_ras.node_insert function in
both the orte_init_stage1 for singleton processes and rds_hostfile.c
rds_base_select.c & rds_base_open.c
- Add and set the global module reference
rds_hostfile.c
- Added functionality to create a new cellid for each hostfile, placing
each entry in the hostfile into the same cellid. Currently this is
commented out with the cellid hard coded to 0, with the intention of
taking this out once ORTE is able to handle multiple cellid's
- Instead of just adding hosts to the Node Segment via a direct call to
the ras_base_node_insert() function. First add the hosts to the Resource
Segment of the GPR using the orte_rds.store_resource() function then use
the API version of orte_ras.node_insert() to store the hosts on the Node
Segment.
- Add 1 new function pointer to module as required by the API.
rds_hostfile_component.c
- Converted this to use the new MCA parameter registration
orte_init_stage1.c
- It is possible that a cellid was not created yet for the current environment.
So I put in some logic to test if the cellid 0 existed. If it does then
continue, otherwise create the cellid so we can properly interact with the
GPR via the RDS.
- For the singleton case we insert some 'dummy' data into the GPR. The RAS
matches this logic, so I took out the duplicate GPR put logic, and
replaced it with a call to the orte_ras.node_insert() function.
- Further before calling orte_ras.node_insert() in the singleton case,
we also call orte_rds.store_resource() to add the singleton node to the
Resource Segment.
Console:
- Added a bunch of new functions. Still experimenting with many aspects of the
implementation. This is a checkpoint, and has very limited functionality.
- Should not be considered stable at the moment.
This commit was SVN r6813.
using orteprobe.
Created a header file for orte_setup_hnp. [HNP = Head Node Process]
General cleanup and added a bit of documentation in orte_setup_hnp.c
Also fixed a cellid tokens issue (circa line 285)
Changed the launched scope from private to public
In orteprobe:
- added reference to orted.h to avoid duplicate header contents in orteprobe.h
- removed the version tag, and put in a verbose argument
- Fixed a buffer packing problem that was causing the parent from receiving the
proper contact information for the new daemon.
This commit was SVN r6802.
If you register a parameter a second time, it overwrites the default
value (this was causing a problem with mpirun not being marked as orte
infrastructure, and therefore thinking that it was a singleton, and
therefore always adding the localhost into the node list).
This commit was SVN r6789.
- converted some things to new MCA param API
- renamed the pls_bproc_seed component struct so its name isn't the same as
the pls_bproc component's struct
- minor bugfixes
This commit was SVN r6774.
his e-mail:
I ran into a small bug in rmaps_rr.c: map_app_by_slot which was
triggered by using multiple app contexts. Basically, if not all the
slots we allocated on a node were used by an app, we would
automatically move onto the next node. This caused a problem with
multiple app contexts when the first app takes a partial allocation of
a node, the second app would not be able to access these slots because
we had already move past the node, and the byslot routine does not
wrap back around the list.
This commit was SVN r6766.
that were set on the command line. This was techinically exactly the
way the code was designed, but it certainly violated the Law of Least
Astonishment (even to its designer ;-) ). So now if you execute
something like this:
mpirun -mca pls_rsh_debug 1 -np 4 hello
You'll see debugging output from the rsh pls component, as you would
expect (this was not previously the case -- the MCA pls_rsh_debug
parame would be set to 1 in the 4 spawned hello processes, but *not*
in the orterun process).
More specifically, MCA parameters will be set in the orterun process
in the following cases:
- The new command line switch "--gmca" (or "-gmca") is used,
indicating that the MCA parameter is "global". --gmca also means
that that MCA parameter will be applied to all context app's. For
example:
mpirun -gmca foo bar -np 1 hello : -np 2 goodbye
The foo MCA param will be set in both the hello and goodbye
processes.
- If there is only one context app. For example:
mpirun -mca pls_rsh_debug 1 -np 4 hello
will set pls_rsh_debug to 1 in both the orterun process and the 4
spawned hello processes.
Also added a few more comments inside orterun to document a somewhat
confusing use of a state variable in a recursive case.
This commit was SVN r6764.
1. dump_xxx - analogous to the registry's dump commands, allows you to examine the contents of the name services' structures
2. get_job_peers - get an array of process names for all processes in the specified job
This commit was SVN r6759.
Somehow, in changing over to the new MCA interfaces, the "set" part of that logic got lost, so the singleton flag was always being set. This should repair some of the anomalous behavior seen recently where the local host was always being used for an application process.
This commit was SVN r6757.
containers when one is requested.
Fix a bug in gpr_replica_del_index_api which doesn't preset num_tokens and
num_keys, but assumes they are 0.
Fix orte_ras_base_node_delete() function to operate properly to delete the
appropriate container in the 'orte-node' segment when requested.
This commit was SVN r6756.
- convert MCA params to the new API
- some style and indenting fixes
- look at local shell, and if [new] MCA param
pls_rsh_assume_same_shell is 1, then assume that the remote shell is
the same as the local shell. If pls_rsh_assume_same_shell is 0, do
a probe to figure out what the remote shell is (NOT CURRENTLY
IMPLEMENTED! you'll get a run-time warning if you set this MCA param
to 0).
- if the remote shell is not csh and not bash, then prefix the remote
command with "( ! [ -e ./.profile ] || . ./.profile;" (and suffix it
with ")") so that we run the .profile on the remote side in order to
set PATHs and the like. See the LAM FAQ for details (will someday
be on the Open MPI FAQ:
http://www.lam-mpi.org/faq/category4.php3#question8)
- add a bunch of debugging output if the MCA param pls_rsh_debug is
enabled (or the top-level debug MCA param is enabled)
- add more help messages (and corresponding calls to opal_show_help())
in help-pls-rsh.txt
This commit was SVN r6731.
- we now properly support multiple application contexts
- much improved error messages, using opal_show_help
- fix some small bugs in the way the processes were discovering their names
- better searching for orted
- use the new mca parameter interface
These changes still need some testing, but they seem stable.
This commit was SVN r6719.
- Add functionality to parse multiple arguments provided in the console
- Cleaned up help function
- Added an option to hide commands from the help menu
Working on launching and reaping of daemons from within the console.
This commit was SVN r6699.
This required a little fiddling with a number of areas. Biggest problem was that it uncovered a potential for an infinite loop to be created in the registry. If a callback function modified the registry, the registry checked the triggers to see if anything had fired. Well, if the original callback was due to a trigger firing, that condition hadn't changed - so the trigger fired again....which caused the callback to be called, which modified the registry, which checked the triggers, etc. etc.
Triggers are now checked and then "flagged" as being "in process" so that the registry will NOT recheck that trigger until all callbacks have been processed. Tried doing this with subscriptions as well, but that caused a problem - when we release processes from a stagegate, they (at the moment) immediately place data on the registry that should cause a subscription to fire. Unfortunately, the system will just hang if that subscription doesn't get processed. So, I have left the subscription system alone - any callback function that modifies the registry in a fashion that will fire a subscription will indeed fire that subscription. We'll have to see if this causes problems - it shouldn't, but a careless user could lock things up if the callback generates a callback to itself.
Also fixed the code that placed a process' RML contact info on the registry to eliminate the leading '/' from the string.
This commit was SVN r6684.
