|
1 | 1 | # -*- text -*-
|
2 | 2 | #
|
3 |
| -# Copyright (c) 2012-2020 Cisco Systems, Inc. All rights reserved. |
| 3 | +# Copyright (c) 2012-2022 Cisco Systems, Inc. All rights reserved. |
4 | 4 | #
|
5 | 5 | # $COPYRIGHT$
|
6 | 6 | #
|
@@ -104,16 +104,6 @@ value will be ignored.
|
104 | 104 | Value: %s
|
105 | 105 | Message: %s
|
106 | 106 | #
|
107 |
| -[device present but not up] |
108 |
| -Open MPI has found a usNIC device that is present / listed in Linux, |
109 |
| -but in a "down" state. It will not be used by this MPI job. |
110 |
| - |
111 |
| -You may wish to check this device, especially if it is unexpectedly |
112 |
| -down. |
113 |
| - |
114 |
| - Local server: %s |
115 |
| - Device name: %s |
116 |
| -# |
117 | 107 | [MTU mismatch]
|
118 | 108 | The MTU does not match on local and remote hosts. All interfaces on
|
119 | 109 | all hosts participating in an MPI job must be configured with the same
|
@@ -224,22 +214,6 @@ Check the resulting "mymap*" files to see the exact pairing of IP
|
224 | 214 | interfaces. Inconsistent results may be indicative of underlying
|
225 | 215 | network misconfigurations.
|
226 | 216 | #
|
227 |
| -[fi_av_insert timeout] |
228 |
| -The usnic BTL failed to create addresses for remote peers within the |
229 |
| -specified timeout. This usually means that ARP requests failed to |
230 |
| -resolve in time. You may be able to solve the problem by increasing |
231 |
| -the usnic BTL's ARP timeout. If that doesn't work, you should |
232 |
| -diagnose why ARP replies are apparently not being delivered in a |
233 |
| -timely manner. |
234 |
| - |
235 |
| -The usNIC interface listed below will be ignored. Your MPI |
236 |
| -application will likely either run with degraded performance and/or |
237 |
| -abort. |
238 |
| - |
239 |
| - Server: %s |
240 |
| - usNIC interface: %s |
241 |
| - Current ARP timeout: %d (btl_usnic_arp_timeout MCA param) |
242 |
| -# |
243 | 217 | [fi_av_eq too small]
|
244 | 218 | The usnic BTL was told to create an address resolution queue that was
|
245 | 219 | too small via the mca_btl_usnic_av_eq_num MCA parameter. This
|
@@ -283,56 +257,6 @@ connectivity map file will not be written.
|
283 | 257 | Working directory: %s
|
284 | 258 | Error: %s (%d)
|
285 | 259 | #
|
286 |
| -[received too many short packets] |
287 |
| -WARNING: The usnic BTL received a significant number of abnormally |
288 |
| -short packets on a single network interface. This may be due to |
289 |
| -corruption or congestion in the network fabric. It may be useful to |
290 |
| -run a physical/layer 0 diagnostic. |
291 |
| - |
292 |
| -Your job will continue, but if this poor network behavior continues, |
293 |
| -you may experience lower-than-expected performance due to overheads |
294 |
| -caused by higher-than-usual retransmission rates (to compensate for |
295 |
| -the corrupted received packets). |
296 |
| - |
297 |
| - Local server: %s |
298 |
| - usNIC interface: %s |
299 |
| - # of short packets |
300 |
| - received so far: %d |
301 |
| - |
302 |
| -You will only receive this warning once per MPI process per job. |
303 |
| - |
304 |
| -If you know that your network environment is lossy/heavily congested |
305 |
| -such that short/corrupted packets are expected, you can disable this |
306 |
| -warning by setting the btl_usnic_max_short_packets MCA parameter to 0. |
307 |
| -# |
308 |
| -[non-receive completion error] |
309 |
| -WARNING: The usnic BTL has detected an error in the completion of a |
310 |
| -non-receive event. This is highly unusual, and may indicate an error |
311 |
| -in the usNIC subsystem on this server. |
312 |
| - |
313 |
| -Your MPI job will continue, but you should monitor the job and ensure |
314 |
| -that it behaves correctly. |
315 |
| - |
316 |
| - Local server: %s |
317 |
| - usNIC interface: %s |
318 |
| - Channel index: %d |
319 |
| - Completion status: %s (%d) |
320 |
| - Work request ID: %p |
321 |
| - Opcode: %s (%d) |
322 |
| - |
323 |
| -If this error keeps happening, you should contact Cisco technical |
324 |
| -support. |
325 |
| -# |
326 |
| -[device present but not up] |
327 |
| -Open MPI has found a usNIC device that is present / listed in Linux, |
328 |
| -but in a "down" state. It will not be used by this MPI job. |
329 |
| - |
330 |
| -You may wish to check this device, especially if it is unexpectedly |
331 |
| -down. |
332 |
| - |
333 |
| - Local server: %s |
334 |
| - Device name: %s |
335 |
| -# |
336 | 260 | [transport mismatch]
|
337 | 261 | Open MPI has found two servers with different underlying usNIC
|
338 | 262 | transports. This is an unsupported configuration; all usNIC devices
|
|
0 commit comments