LAM/MPI logo

LAM/MPI General User's Mailing List Archives

  |   Home   |   Download   |   Documentation   |   FAQ   |   all just in this list

From: Sam Steingold (sds_at_[hidden])
Date: 2007-08-16 13:13:13


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Sam Steingold wrote:
> what does the error "invalid tag argument" mean and how do I avoid it?

more information:
I use OCamlMPI (http://caml.inria.fr/cgi-bin/hump.en.cgi?contrib=401,
http://cristal.inria.fr/~xleroy/software.html#ocamlmpi).

the only tags that we pass around are "0" and "1".
alas, every now and then I see this in my logs:
"received tag=-32766 from rank=-32766"
(and a crash when I try to pass this invalid tag to the return message).
instead of the normal "received tag=0 from rank=0" for clients and
"received tag=0 from rank=<node_num>" for the server.
grep|sed|sort|uniq -c|sort:
     14 received tag=-32766 from rank=-32766
     17 received tag=0 from rank=4
     49 received tag=0 from rank=88
     59 received tag=0 from rank=59
     60 received tag=0 from rank=149
     75 received tag=0 from rank=5
     93 received tag=0 from rank=72
    101 received tag=0 from rank=89
    108 received tag=0 from rank=113
    110 received tag=0 from rank=14
    110 received tag=0 from rank=90
    112 received tag=0 from rank=133
    125 received tag=0 from rank=137
    141 received tag=0 from rank=62
    159 received tag=0 from rank=104
    163 received tag=0 from rank=134
    164 received tag=0 from rank=170
    167 received tag=0 from rank=109
    168 received tag=0 from rank=22
    171 received tag=0 from rank=177
    178 received tag=0 from rank=71
    184 received tag=0 from rank=187
    186 received tag=0 from rank=92
    188 received tag=0 from rank=186
    195 received tag=0 from rank=45
    196 received tag=0 from rank=180
    197 received tag=0 from rank=54
    204 received tag=0 from rank=164
    208 received tag=0 from rank=110
    211 received tag=0 from rank=70
    214 received tag=0 from rank=139
    221 received tag=0 from rank=91
    225 received tag=0 from rank=50
    226 received tag=0 from rank=78
    228 received tag=0 from rank=116
    229 received tag=0 from rank=123
    229 received tag=0 from rank=84
    230 received tag=0 from rank=160
    230 received tag=0 from rank=42
    231 received tag=0 from rank=67
    232 received tag=0 from rank=141
    232 received tag=0 from rank=95
    233 received tag=0 from rank=68
    238 received tag=0 from rank=2
    240 received tag=0 from rank=1
    240 received tag=0 from rank=106
    245 received tag=0 from rank=51
    247 received tag=0 from rank=165
    247 received tag=0 from rank=188
    247 received tag=0 from rank=26
    247 received tag=0 from rank=32
    249 received tag=0 from rank=183
    252 received tag=0 from rank=93
    256 received tag=0 from rank=128
    260 received tag=0 from rank=77
    263 received tag=0 from rank=100
    263 received tag=0 from rank=179
    264 received tag=0 from rank=131
    269 received tag=0 from rank=125
    269 received tag=0 from rank=174
    269 received tag=0 from rank=28
    273 received tag=0 from rank=152
    274 received tag=0 from rank=25
    276 received tag=0 from rank=176
    276 received tag=0 from rank=86
    277 received tag=0 from rank=148
    277 received tag=0 from rank=3
    277 received tag=0 from rank=44
    278 received tag=0 from rank=153
    279 received tag=0 from rank=87
    282 received tag=0 from rank=48
    283 received tag=0 from rank=175
    285 received tag=0 from rank=126
    292 received tag=0 from rank=98
    294 received tag=0 from rank=17
    295 received tag=0 from rank=173
    296 received tag=0 from rank=8
    298 received tag=0 from rank=129
    299 received tag=0 from rank=118
    299 received tag=0 from rank=163
    300 received tag=0 from rank=132
    302 received tag=0 from rank=107
    302 received tag=0 from rank=85
    304 received tag=0 from rank=18
    306 received tag=0 from rank=16
    307 received tag=0 from rank=120
    307 received tag=0 from rank=127
    308 received tag=0 from rank=105
    309 received tag=0 from rank=150
    310 received tag=0 from rank=130
    310 received tag=0 from rank=52
    312 received tag=0 from rank=178
    312 received tag=0 from rank=66
    313 received tag=0 from rank=39
    317 received tag=0 from rank=115
    318 received tag=0 from rank=74
    319 received tag=0 from rank=56
    320 received tag=0 from rank=161
    320 received tag=0 from rank=34
    320 received tag=0 from rank=64
    321 received tag=0 from rank=49
    322 received tag=0 from rank=145
    323 received tag=0 from rank=114
    325 received tag=0 from rank=69
    325 received tag=0 from rank=73
    326 received tag=0 from rank=13
    326 received tag=0 from rank=167
    326 received tag=0 from rank=181
    326 received tag=0 from rank=182
    326 received tag=0 from rank=30
    328 received tag=0 from rank=81
    329 received tag=0 from rank=101
    329 received tag=0 from rank=79
    331 received tag=0 from rank=124
    333 received tag=0 from rank=162
    335 received tag=0 from rank=23
    336 received tag=0 from rank=75
    340 received tag=0 from rank=156
    340 received tag=0 from rank=58
    342 received tag=0 from rank=121
    343 received tag=0 from rank=35
    344 received tag=0 from rank=166
    346 received tag=0 from rank=185
    350 received tag=0 from rank=103
    351 received tag=0 from rank=40
    355 received tag=0 from rank=117
    356 received tag=0 from rank=99
    357 received tag=0 from rank=122
    358 received tag=0 from rank=184
    360 received tag=0 from rank=83
    361 received tag=0 from rank=6
    366 received tag=0 from rank=143
    366 received tag=0 from rank=31
    367 received tag=0 from rank=111
    367 received tag=0 from rank=135
    367 received tag=0 from rank=24
    370 received tag=0 from rank=20
    370 received tag=0 from rank=53
    375 received tag=0 from rank=102
    375 received tag=0 from rank=169
    376 received tag=0 from rank=171
    381 received tag=0 from rank=112
    381 received tag=0 from rank=80
    382 received tag=0 from rank=146
    382 received tag=0 from rank=46
    383 received tag=0 from rank=65
    385 received tag=0 from rank=97
    386 received tag=0 from rank=168
    390 received tag=0 from rank=37
    390 received tag=0 from rank=76
    393 received tag=0 from rank=119
    393 received tag=0 from rank=151
    393 received tag=0 from rank=159
    396 received tag=0 from rank=154
    398 received tag=0 from rank=60
    399 received tag=0 from rank=57
    400 received tag=0 from rank=19
    400 received tag=0 from rank=9
    405 received tag=0 from rank=142
    405 received tag=0 from rank=147
    409 received tag=0 from rank=10
    409 received tag=0 from rank=41
    411 received tag=0 from rank=55
    415 received tag=0 from rank=15
    415 received tag=0 from rank=82
    418 received tag=0 from rank=11
    419 received tag=0 from rank=172
    420 received tag=0 from rank=36
    421 received tag=0 from rank=38
    422 received tag=0 from rank=63
    424 received tag=0 from rank=108
    426 received tag=0 from rank=158
    428 received tag=0 from rank=7
    435 received tag=0 from rank=12
    437 received tag=0 from rank=33
    442 received tag=0 from rank=144
    443 received tag=0 from rank=29
    443 received tag=0 from rank=94
    445 received tag=0 from rank=138
    452 received tag=0 from rank=43
    455 received tag=0 from rank=61
    456 received tag=0 from rank=21
    463 received tag=0 from rank=136
    477 received tag=0 from rank=140
    478 received tag=0 from rank=27
    484 received tag=0 from rank=47
    495 received tag=0 from rank=96
    497 received tag=0 from rank=157
    512 received tag=0 from rank=155
  57899 received tag=0 from rank=0

so, is there a better explanation than "network hardware noise generated
by gremlings"?

and what should I do?
request a retransmit?
I need to drop the current message, right?
how? (since the tag came corrupted, it is likely that the message is
corrupted too, so I cannot just read the expected number of bytes and
then drop them).

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGxIWpPp1Qsf2qnMcRApZ2AJ9XUeblv8bM9RFR2GbT4pLYYPrFUACfZhHl
7adUwOQ5IlqGkj6ZejE8vn8=
=xWsw
-----END PGP SIGNATURE-----