LAM/MPI logo

LAM/MPI Development Mailing List Archives

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

From: Martin Knoblauch (spamtrap_at_[hidden])
Date: 2006-06-09 02:08:49


>
> Hi,
>
> one of my employers customers had a problem with a LAM (7.0.4) based
> application throwing a "no-localhost" error on lamboot from an
> apparently perfect bhosts configuration. After some confusion and
> guessing we found that SuSE had put a line:
>
> 127.0.0.2 hostname.domain hostname
>
> in /etc/hosts (IPV6 stuff ??). After removing it everything was OK
> again. So far just a user error.
>
> BUT: wouldn't it be better if lamboot would throw a
> "resolve-localhost" instead? It would have definitely been more
> meaningfull and helpfull, at least in my opinion.
>
> Maybe "lam_ssi_boot_rsh_verify_nodes" should be modified to match
> any
> 127.0.0.X address for returning "resolve-localhost"? Maybe something
> like:
>
> --- ssi_boot_rsh.c-orig 2006-06-08 13:33:00.000000000 +0200
> +++ ssi_boot_rsh.c 2006-06-08 13:40:56.000000000 +0200
> @@ -357,12 +357,12 @@
>
> /* If there is more than one node in the universe -- and we are
> not
> using the -l option -- ensure that the origin node does not
> - resolve to 127.0.0.1, or the remote nodes will NOT be sent info
> + resolve to 127.0.0.X, or the remote nodes will NOT be sent info
> back to lamboot */
>
> if (nnodes > 1 && !ao_taken(lam_ssi_boot_optd, "l")) {
> - if (strcmp("127.0.0.1",
> - inet_ntoa(nodes[origin].lnd_addr.sin_addr)) == 0) {
> + if (strcmp(inet_ntoa(nodes[origin].lnd_addr.sin_addr,
> + "127.0.0")) != NULL) {
> show_help("boot-ssi", "resolve-localhost",
> "The boot SSI rsh module", nodes[origin].lnd_hname);
> return LAMERROR;
>
> Just ignore me if I talk total nonsense.
>

 which of course I did... :-) I wanted to use "strstr" instead of
"strcmp".

diff -u ssi_boot_rsh.c-orig ssi_boot_rsh.c
--- ssi_boot_rsh.c-orig 2006-06-08 13:33:00.000000000 +0200
+++ ssi_boot_rsh.c 2006-06-09 08:06:09.000000000 +0200
@@ -357,12 +357,12 @@

   /* If there is more than one node in the universe -- and we are not
      using the -l option -- ensure that the origin node does not
- resolve to 127.0.0.1, or the remote nodes will NOT be sent info
+ resolve to 127.0.0.X, or the remote nodes will NOT be sent info
      back to lamboot */

   if (nnodes > 1 && !ao_taken(lam_ssi_boot_optd, "l")) {
- if (strcmp("127.0.0.1",
- inet_ntoa(nodes[origin].lnd_addr.sin_addr)) == 0) {
+ if (strstr(inet_ntoa(nodes[origin].lnd_addr.sin_addr,
+ "127.0.0")) != NULL) {
       show_help("boot-ssi", "resolve-localhost",
                "The boot SSI rsh module", nodes[origin].lnd_hname);
       return LAMERROR;

------------------------------------------------------
Martin Knoblauch
email: k n o b i AT knobisoft DOT de
www: http://www.knobisoft.de