1
1
Граф коммитов

300 Коммитов

Автор SHA1 Сообщение Дата
Ralph Castain
22dc05194e Minor cleanup - explicitly NULL the last member of a function pointer module. Should default to that anyway, but this is cosmetically nicer. 2015-10-30 08:19:55 -07:00
Ralph Castain
267ca8fcd3 Cleanup the PMIx direct modex support. Add an MCA parameter pmix_base_async_modex that will cause the async modex to be used when set to 1. Default it to 0 for now
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.
2015-10-27 17:31:56 -07:00
Ralph Castain
8f6855459d Cleanup some coverity warnings 2015-09-30 10:33:53 -07:00
Ralph Castain
c1bbbb5e2f Remove the last involvement of the OOB system from the MPI layer, remove the no-longer-needed usock/oob component, and have procs no longer open the RML, OOB, ROUTED, and GRPCOMM frameworks as PMIx now provides all required app-mpirun cmds 2015-09-15 13:08:35 -07:00
Ralph Castain
22d7c0081a Fix the no-disconnect test by resolving a segfault on free - opal_dss.unload will return the remaining unpacked portion of a buffer. As such, it cannot return the pointer to that info as it might be partway inside of a malloc'd region. So copy the data out of the buffer. 2015-09-11 13:01:35 -07:00
Ralph Castain
dc5796b8a1 Revert "Revert "Fix the handling of cpusets so we get the correct cpuset for each local peer. Add the ability to indicate that a modex request is "optional" so we don't call the server if we don't find the value. Take advantage of that to allow the MPI layer to decide that the lack of locality info indicates non-local""
Fix the locality computation by correctly computing the vpid of the local peer

This reverts commit open-mpi/ompi@6a8fad49e5.
2015-09-11 08:29:51 -07:00
Ralph Castain
6a8fad49e5 Revert "Fix the handling of cpusets so we get the correct cpuset for each local peer. Add the ability to indicate that a modex request is "optional" so we don't call the server if we don't find the value. Take advantage of that to allow the MPI layer to decide that the lack of locality info indicates non-local"
This reverts commit f94f3cda21.
2015-09-11 02:01:25 -07:00
Ralph Castain
f94f3cda21 Fix the handling of cpusets so we get the correct cpuset for each local peer. Add the ability to indicate that a modex request is "optional" so we don't call the server if we don't find the value. Take advantage of that to allow the MPI layer to decide that the lack of locality info indicates non-local 2015-09-10 10:25:30 -07:00
Ralph Castain
e6add86e4f Deal with connect/accept between two jobs from different mpirun's. Somewhat optimize connect/accept by using MPI bcast to distribute the participants instead of another PMIx lookup. Cleanup some Coverity issues. 2015-09-07 09:19:24 -07:00
Ralph Castain
37c3ed68e7 Cleanup connect/disconnect and bring comm_spawn back online! 2015-09-06 10:27:39 -07:00
rhc54
665b30376a Merge pull request #868 from rhc54/topic/hwloc
Remove OPAL_HAVE_HWLOC qualifier and error out if --without-hwloc is given
2015-09-04 17:58:07 -07:00
Ralph Castain
d97bc29102 Remove OPAL_HAVE_HWLOC qualifier and error out if --without-hwloc is given 2015-09-04 16:54:40 -07:00
Ralph Castain
f6948c2bb4 Sync with PMIx master 43e45c3. Get multi-node publish/lookup/unpublish working 2015-09-04 10:07:17 -07:00
Ralph Castain
a772b46c15 Bring the MPI_Publish and friends online 2015-09-02 12:04:07 -07:00
Ralph Castain
38ba54366c Fix shared memory operations by resolving local peers 2015-08-30 12:07:14 -07:00
Ralph Castain
0d5814b5ca Cleanup Coverity issues 2015-08-29 21:19:27 -07:00
Ralph Castain
cf6137b530 Integrate PMIx 1.0 with OMPI.
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
2015-08-29 16:04:10 -07:00
Ralph Castain
0b1d4b62be Cleanup some cruft and update to coordinate with CM operations:
* 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
2015-08-12 10:32:14 -07:00
Ralph Castain
9bc384282a Fix an annoying segfault caused by incorrect indentation in a loop that causes the buffer to not be created prior to packing. 2015-08-01 10:01:47 -07:00
Nathan Hjelm
4d92c9989e more c99 updates
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>
2015-06-25 10:14:13 -06:00
Ralph Castain
869041f770 Purge whitespace from the repo 2015-06-23 20:59:57 -07:00
Gilles Gouaillardet
ec679b3fc2 orte/orted: fix misc memory leaks 2015-06-17 11:17:55 +09:00
Gilles Gouaillardet
27b4727fcf orte/orted: fix misc memory leak
as reported by Coverity with CID 743448
2015-06-17 11:17:55 +09:00
Ralph Castain
ea35e47228 Fat SMPs (i.e., systems with nodes containing large numbers of cpus) were failing to start due to connection failures of the opal/pmix support. Root cause was that (a) we were setting the client socket to non-blocking before calling connect, and (b) the server was using the event library to harvest the accepts, and also did the handshake while in that event. So the server would backup beyond the connection backlog limit, and we would fail.
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.
2015-05-29 14:37:14 -07:00
Ralph Castain
b5382c9bf9 Rework the OOB selection logic to allow a component (e.g., usock) to direct that it be the sole active component. Remove prior disqualifying code in the oob/tcp component as it was too restrictive - if usock wasn't able to run, it left apps with no way to communicate to their daemon. Have the local daemon check the global modex for the RML URI info of the local procs so it can route messages between them when tcp is the primary channel.
A few other minor cleanups included.
2015-05-08 11:15:21 -07:00
Gilles Gouaillardet
2e384a3b65 initialize common symbols from orte
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
2015-05-08 10:11:58 +09:00
rhc54
bc016617a0 Merge pull request #501 from rhc54/topic/sec2
Support authentication across security domains
2015-03-30 09:59:43 -07:00
Nadezhda Kogteva
a828eada98 sm dstore: set pmix segment size to proper value 2015-03-30 13:34:25 +03:00
Ralph Castain
d07dc362d5 Ensure we can authenticate when crossing security domains by including all available credentials, and letting the receiver use the highest priority one they have in common. 2015-03-28 20:34:26 -07:00
Ralph Castain
1b24536941 Allow for different security domains. Let the initiator of the connection determine the method to be used - if the receiver cannot support it, then that's an error that will cause the connection attempt to fail. 2015-03-25 13:22:01 -07:00
Ralph Castain
43a3baad5e Ensure we use the first compute node's topology for mapping
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
2015-03-19 16:30:36 -07:00
Ralph Castain
d7d8ae46ed We no longer pass the RML URI for procs launched via mpirun as the daemon has no need for that info. 2015-03-17 06:10:20 -07:00
Gilles Gouaillardet
ab78c7f54a orted/pmix: fix misc resource leak
as reported by Coverity with CID 1269844
2015-02-27 19:25:55 +09:00
Ralph Castain
3de8c5c7c6 Cleanup the munge support - the credential cannot be reused for multiple connections 2015-02-10 20:34:35 -08:00
Ralph Castain
063e4c9989 Cleanup the pretty-print of odls cmds as some were missing. Add a new cmd to terminate the DVM, which the HNP will use to trun around and issue an xcast to the DVM. 2015-02-10 08:27:13 -08:00
Ralph Castain
3ae3b96c17 Fix master compilation - a buried header dependency must have been removed. 2015-02-10 07:22:10 -08:00
Ralph Castain
3478def791 Ensure that nodes get included in the nidmap when spawning a new DVM job - we really only need to do this once, but for now we do it for every job until we work out how to avoid the duplication. Remove debug from orte-dvm tool 2015-02-09 23:47:46 -05:00
Ralph Castain
251084a2da When a tool requests the spawn of a new job, then exclusively forward output to that tool - the DVM should not output its own copy as well. 2015-02-04 07:59:47 -08:00
Ralph Castain
2b0b012460 Continue refinement of the DVM operations. Send the spawn request to the right place (it helps) as it isn't a comm_spawn request and has to be treated a little differently. Ensure IO gets forwarded back to the tool. Ensure the tool outputs show_help locally as there is no place to send it. 2015-02-04 06:21:54 -08:00
Ralph Castain
fcec24b2a4 Minor cleanups to handle comm_spawn and singletons 2015-01-27 09:29:42 -06:00
Nadezhda Kogteva
05af80b302 Fix commit bffb2b7a4b which broke pmix server functionality 2014-12-24 13:25:23 +02:00
Ralph Castain
2581b41d08 Continue refactoring code by splitting the msg processing from the sendrecv code 2014-12-17 19:57:14 -08:00
Ralph Castain
f489e871c2 Take first step towards refactoring the PMIx server code by splitting out the proc_map function into its own file. Update ignore to include .DS_Store from the Mac 2014-12-17 19:08:52 -08:00
Ralph Castain
573a574a3c Remove an unused dstore type that was redundant with another one. Define a corresponding PMIX_NODE_ID type (contains the vpid of the daemon hosting the proc) and ensure that the PMIx server includes that info in its process map 2014-12-15 12:11:13 -08:00
Ralph Castain
a22cc45769 Close the pmix server sockets on exec 2014-12-13 20:30:21 -08:00
Ralph Castain
9658256a98 Restore the passing of the complete job map to the local proc on first get_attr so the info can be used by the MPI layer without continual calls back to the server. We'll find a more memory efficient method later. 2014-12-13 18:44:09 -08:00
Ralph Castain
bffb2b7a4b Correct some issues with variables used before being set 2014-12-12 17:23:32 -08:00
Ralph Castain
bb529ebd8e Revise the way we handle hetero nodes as users are finding this (a) a significant surprise, and (b) confusing as to when it is required. So try to automate it a bit by creating a topology "signature" that mpirun can share on the cmd line with the remote daemons, thus allowing them to check to see if they match. This isn't comprehensive of course - for now, it only checks the number of each type of hwloc object on the node. This is good enough to pickup major differences (e.g., where we have different numbers of sockets or assigned core bindings).
Retain the hetero-nodes flag for those cases where the user *knows* that there are differences and our automated system isn't good enough to see it.

Will obviously require further refinement as we find out which variances it can detect, and which it cannot.
2014-12-08 15:38:14 -08:00
Ralph Castain
c88f181efe Fix singleton comm-spawn, yet again. The new grpcomm collectives require a complete knowledge of every active proc in the system in case they participate in a collective. So ensure we pass the required job info when we spawn new daemons, and construct the necessary connections to allow grpcomm to operate. 2014-12-03 18:11:17 -08:00
Ralph Castain
f48b9012cb Some minor cleanup. We really don't need another peer error constant to indicate that a peer closed as we already have one for "connection failed", and that's all we really know. Update the orte constants to track their opal equivalents. 2014-11-25 08:02:29 -08:00