Re: Re: [Queue-developers] new intermediate development Queue version

[Gert V. d. E. <gvd...@sc...>]

Dear QingLong,

>  > How do I work around this ? 
>  >
>     It looks like you are using development versions of the tools.
>   If so, then you are on your own here, I am not able to help you,
>   so you would probably have to try yourself to find out if this is
>   autoconf/automake or Queue bug. And if you find that it is Queue
>   configure.in that should be fixed, please, teach us how should we
>   modify it to meet autoconf requirements. Thank you.

I'm sorry, I'm a coward under time pressure to get a queueing system working on our cluster and a non-expert on autoconf/automake, so I grabbed and installed the versions you mentioned. All compilation went well after that.

I've done some playing around, these are my observations:

- compiled queue and queued with --enable-root (no --enable-manager)
- queued started with --debug and -t 10 on 5 hosts, each maxexec 2
- submitted 12 jobs for the now queue (using -i -w -p)
- 10 start immediately, as expected
- 1 is on hold, as expected
- last one gives me back the shell after some seconds with the announcement 'Alarm clock'

- queued started with --debug --foreground -t 10 on 5 hosts, each maxexec 1
- submitted 7 jobs for the now queue (using -v -i -w -p)
 -5 start immediately, as expected
- 2 are on hold
- I get lots of messages from queued among which there are 

file now/CFDIR/cfm701151391 has 0 length: Bad file descriptor
now/CFDIR/cfm701151391 has 0 length.
SENDMAIL: To 'root' from 'root': Subject: queued error on bohr: file now/CFDIR/cfm701151391 has 0 length: Interrupted system call

Requesting load average for queue "now" on host "pauli"...
queue_persistent_connect(): connect()ing to 192.168.1.5:1423 ...
queue_persistent_connect(): connect()ed  to 192.168.1.5:1423 on socket 6.
queue_reliable_fread(): select()ing on socket 6...
queue_reliable_fread(): select(6) timed out.
getrldavg(): failed to fread() from stream opened on socket 6.
getrldavg(): close(6).
getrldavg(): ### failed to get load from pauli
	     ### returning 1.00e+08 as rejection designator.

Are these messages something you would expect from queued or are they indicating that something is wrong? 

If you need more specific debugging information (or if you have a test-case I can run on our system), tell me what to look for....


Have a nice day,

Gert