Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Ashraf_Akkode
Beginner
114 Views

sh: line 1: 26052 Alarm clock /opt/intel/impi/4.0.0.028/intel64/bin

Dear Sir,

When I try to start mpdboot using the following way I have been receiving the following error.
 

***************************************************
[g201002860@hpc081 ~]$ mpdboot -v -n 4 -r ssh -f mpd.hosts
sh: line 1: 26052 Alarm clock /opt/intel/impi/4.0.0.028/intel64/bin/mpdallexit.py > /dev/null
running mpdallexit on hpc081
LAUNCHED mpd on hpc081 via
mpdboot_hpc081 (recv_dict_msg 731):recv_dict_msg raised exception: errmsg=::
mpdtb:
/opt/intel/impi/4.0.0.028/intel64/bin/mpdlib.py, 731, recv_dict_msg
, 890, handle_mpd_output
, 595, mpdboot
, 973, ?

mpdboot_hpc081 (handle_mpd_output 892): failed to ping mpd on hpc081; received output={}

***************************************

Coud you please help to me to resolve this issue.
Ashraf
0 Kudos
3 Replies
Dmitry_K_Intel2
Employee
114 Views

Hi Ashraf,

Did you set up password-less ssh? Can you run 'hostname' command on any node?
[g201002860@hpc081 ~]$ssh hpc082 hostname

This is the first requirement.

Regards!
Dmitry
Ashraf_Akkode
Beginner
114 Views


Dear Dmity

Thanks for your comment
It is resolved.It was due to multiple mpdboot deamons are running.

Regards
Ashraf
Adhip_A_
Beginner
114 Views

Although coming after two years, but for those, who might end up here, after searching online. 

'Alarm clock' error might be displayed if a mpd process crashes badly.  Then execute

mpdcleanup

This command cleans up the crashed files and allows mpd to run again. This might need to be done on each node. Instead one could also use

mpdcleanup -a

for cleaning up all the nodes at once.

regards,

Adhip

Reply