I forgot to post that I found the code in print_bad_nodes in gexec.c:


static void print_bad_nodes(int *bad_nodes, char **ips)
{
    char    host[1024];
    char    *ip;
    int     bad_node;
    int     bad_nodes_copy[GEXEC_MAX_NPROCS];

    bitmask_copy(bad_nodes_copy, bad_nodes, GEXEC_MAX_NPROCS);
    while ((bad_node = bitmask_ffs(bad_nodes_copy, GEXEC_MAX_NPROCS)) != -1) {
        ip = ips[bad_node];
        net_iptohost(ip, host, 1024);
        printf("Could not connect to %s (%s)\n", host, ip);
        bitmask_clr(bad_nodes_copy, GEXEC_MAX_NPROCS, bad_node);
    }
}


Is this routine called for ALL nodes, and then the error is
trapped?  Or is this routine ONLY called for bad nodes?

In other words, should I look at bitmask_ffs or should
I look why this was called in the first place?  I think
it is always called.

Thanks in advance,
Joe


Joe Griffin wrote:
Hello,

As you can see, I am a newbie to Ganglia.  I installed:

authd, libe, ganglia-monitor-core-lib, and gexec.

The error I am getting is:

virtue:/scratch/gang/gexec-0.3.3# gexec -n 0 uname
Could not connect to virtue (192.168.3.1)

I can "ssh" though:
virtue:/scratch/gang/gexec-0.3.3# ssh virtue date
Thu May  2 11:18:21 PDT 2002

Can someone please tell me what I need to do?

I am not sure what to set GEXEC_SVRS to.
I have a 5 node cluster (headnode:virtue)
other nodes msc[1-4].

I tried setting GEXEC_SVRS to "virtue" and "msc1"
but got the same error.  I also looked in the
archives:
https://sourceforge.net/mailarchive/forum.php?max_rows=25&style=ultimate&offset=25&forum_id=7186

But did not see anything.

When I rebuilt gexec, I tried w/ and w/o --enable-ganglia.

I am running on:
Pentium III's
kernel: 2.4.6
glibc 2.2.3
openssh 3.1
openssl 0.9.6c
libe 0.2.0
authd 0.2.0
ganglia-monitor-core-lib 2.3.0
gexec 0.3.3

In summary, ssh works, but gexec does not.
I did all of the openssl type commands:
virtue:/scratch/gang# ls -l /etc/auth*
-rw-------    1 root     root          497 May  1 12:47 /etc/auth_priv.pem
-rw-r--r--    1 root     root          182 May  1 12:47 /etc/auth_pub.pem


Thanks,
Joe


_______________________________________________________________

Have big pipes? SourceForge.net is looking for download mirrors. We supply
the hardware. You get the recognition. Email Us: [EMAIL PROTECTED]
_______________________________________________
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general






Reply via email to