[mvapich-discuss] [BUG] Wrong error message with BLCR

[Maksym Planeta]

Hello,

I found another bug with BLCR. When checkpoint support is not compiled 
in, but the user still specifies arguments related to checkpoints, hydra 
prints following error message, when checkpoint creation is attempted:

$ mpiexec -ckpoint-interval 10 -ckpoint-prefix /home/planeta/opt/chkpt 
-np 4 -hosts 172.31.128.50,172.31.128.51 ./bin/lu.B.4
...
[proxy:0:1 at os-dhcp017] requesting checkpoint
[proxy:0:1 at os-dhcp017] checkpoint completed
[proxy:0:0 at planeta-ib1] requesting checkpoint
[proxy:0:0 at planeta-ib1] checkpoint completed
[proxy:0:1 at os-dhcp017] HYDT_ckpoint_blcr_checkpoint 
(tools/ckpoint/blcr/ckpoint_blcr.c:241): Checkpointing failed.  Make 
sure BLCR kernel module is loaded. Unknown error 2356
[proxy:0:1 at os-dhcp017] ckpoint_thread (tools/ckpoint/ckpoint.c:76): blcr 
checkpoint returned error
[proxy:0:0 at planeta-ib1] HYDT_ckpoint_blcr_checkpoint 
(tools/ckpoint/blcr/ckpoint_blcr.c:241): Checkpointing failed.  Make 
sure BLCR kernel module is loaded. Unknown error 2356
[proxy:0:0 at planeta-ib1] ckpoint_thread (tools/ckpoint/ckpoint.c:76): 
blcr checkpoint returned error

Hydra complains that there is no BLCR module found, but in fact it 
should either complain on wrong arguments or on the absence of 
checkpoint support.

$ mpiname -a
MVAPICH2 2.1 Fri Apr 03 20:00:00 EDT 2015 ch3:nemesis

Compilation
CC: gcc    -DNDEBUG -DNVALGRIND -O2
CXX: g++   -DNDEBUG -DNVALGRIND -O2
F77: gfortran   -O2
FC: gfortran   -O2

Configuration
--prefix=/home/planeta/opt/apps/mvapich/2.1 --enable-fortran=all 
--with-device=ch3:nemesis:ib

-- 
Regards,
Maksym Planeta

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5154 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://mailman.cse.ohio-state.edu/pipermail/mvapich-discuss/attachments/20150818/3a675dd3/attachment-0001.p7s>