You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-user@hadoop.apache.org by Andrew McNabb <am...@mcnabbs.org> on 2006/10/17 21:19:45 UTC
Networking problem in 0.7.1
I just upgraded from Hadoop 0.6.2 to 0.7.1. I'm having a new networking
problem now. My mapred.job.tracker now only listens on localhost
(127.0.0.1). All of the other services seem to be behaving correctly.
In hadoop-site.xml, I have mapred.job.tracker set to "prodigy:50006" and
fs.default.name to "prodigy:50005". I have not overridden
mapred.job.tracker.info.bindAddress anywhere, so I would assume it's
still at 0.0.0.0 as in hadoop-default.xml. Here's what "lsof -i |grep
^java" gives me:
java 18233 amcnabb 3u IPv4 63875 TCP prodigy.aml.cs.byu.edu:38083->sage.aml.cs.byu.edu:ldap (ESTABLISHED)
java 18233 amcnabb 8u IPv6 64013 TCP *:50070 (LISTEN)
java 18233 amcnabb 10u IPv6 64014 TCP localhost.localdomain:50005 (LISTEN)
java 18233 amcnabb 13u IPv6 65342 TCP localhost.localdomain:50005->localhost.localdomain:45751 (ESTABLISHED)
java 18311 amcnabb 6u IPv6 64091 TCP *:50075 (LISTEN)
java 18311 amcnabb 10u IPv6 65341 TCP localhost.localdomain:45751->localhost.localdomain:50005 (ESTABLISHED)
java 18311 amcnabb 12u IPv6 64047 TCP *:50010 (LISTEN)
java 18377 amcnabb 3u IPv4 64087 TCP prodigy.aml.cs.byu.edu:38097->sage.aml.cs.byu.edu:ldap (ESTABLISHED)
java 18377 amcnabb 9u IPv6 64835 TCP localhost.localdomain:50006 (LISTEN)
java 18377 amcnabb 12u IPv6 64838 TCP *:50030 (LISTEN)
java 18459 amcnabb 7u IPv6 64252 TCP *:50060 (LISTEN)
java 18459 amcnabb 10u IPv6 64255 TCP *:50050 (LISTEN)
For some reason, the fs and jobtracker are listening on
localhost.localdomain instead of *. Any ideas on what could be going
wrong? Also, is there a standard port number to use for the
fs.default.name, mapred.job.tracker, etc. (I just made them up)?
--
Andrew McNabb
http://www.mcnabbs.org/andrew/
PGP Fingerprint: 8A17 B57C 6879 1863 DE55 8012 AB4D 6098 8826 6868
Re: Networking problem in 0.7.1
Posted by Andrew McNabb <am...@mcnabbs.org>.
On Tue, Oct 17, 2006 at 01:14:18PM -0700, Robert Firby wrote:
> I had a problem just like this when I upgraded too. It fixed my problem to
> change the machine names in the hadoop.site.xml file to be fully qualified.
>
> Ie. Change prodigy:50006 to be prodigy.whatever.com:50006
>
Weird. I'll go try that. Have you submitted a bug report yet?
--
Andrew McNabb
http://www.mcnabbs.org/andrew/
PGP Fingerprint: 8A17 B57C 6879 1863 DE55 8012 AB4D 6098 8826 6868
Re: Networking problem in 0.7.1
Posted by Andrew McNabb <am...@mcnabbs.org>.
On Tue, Oct 17, 2006 at 01:14:18PM -0700, Robert Firby wrote:
> I had a problem just like this when I upgraded too. It fixed my problem to
> change the machine names in the hadoop.site.xml file to be fully qualified.
>
> Ie. Change prodigy:50006 to be prodigy.whatever.com:50006
>
It turns out that all I had to do was set the FQDN to
prodigy.aml.cs.byu.edu. Apparently Hadoop is doing something like a
"hostname -f" to decide what address to bind on.
--
Andrew McNabb
http://www.mcnabbs.org/andrew/
PGP Fingerprint: 8A17 B57C 6879 1863 DE55 8012 AB4D 6098 8826 6868
Re: Networking problem in 0.7.1
Posted by Robert Firby <ji...@gmail.com>.
I had a problem just like this when I upgraded too. It fixed my problem to
change the machine names in the hadoop.site.xml file to be fully qualified.
Ie. Change prodigy:50006 to be prodigy.whatever.com:50006
Jim
On 10/17/06 12:19 PM, "Andrew McNabb" <am...@mcnabbs.org> wrote:
> I just upgraded from Hadoop 0.6.2 to 0.7.1. I'm having a new networking
> problem now. My mapred.job.tracker now only listens on localhost
> (127.0.0.1). All of the other services seem to be behaving correctly.
> In hadoop-site.xml, I have mapred.job.tracker set to "prodigy:50006" and
> fs.default.name to "prodigy:50005". I have not overridden
> mapred.job.tracker.info.bindAddress anywhere, so I would assume it's
> still at 0.0.0.0 as in hadoop-default.xml. Here's what "lsof -i |grep
> ^java" gives me:
>
> java 18233 amcnabb 3u IPv4 63875 TCP
> prodigy.aml.cs.byu.edu:38083->sage.aml.cs.byu.edu:ldap (ESTABLISHED)
> java 18233 amcnabb 8u IPv6 64013 TCP *:50070 (LISTEN)
> java 18233 amcnabb 10u IPv6 64014 TCP
> localhost.localdomain:50005 (LISTEN)
> java 18233 amcnabb 13u IPv6 65342 TCP
> localhost.localdomain:50005->localhost.localdomain:45751 (ESTABLISHED)
> java 18311 amcnabb 6u IPv6 64091 TCP *:50075 (LISTEN)
> java 18311 amcnabb 10u IPv6 65341 TCP
> localhost.localdomain:45751->localhost.localdomain:50005 (ESTABLISHED)
> java 18311 amcnabb 12u IPv6 64047 TCP *:50010 (LISTEN)
> java 18377 amcnabb 3u IPv4 64087 TCP
> prodigy.aml.cs.byu.edu:38097->sage.aml.cs.byu.edu:ldap (ESTABLISHED)
> java 18377 amcnabb 9u IPv6 64835 TCP
> localhost.localdomain:50006 (LISTEN)
> java 18377 amcnabb 12u IPv6 64838 TCP *:50030 (LISTEN)
> java 18459 amcnabb 7u IPv6 64252 TCP *:50060 (LISTEN)
> java 18459 amcnabb 10u IPv6 64255 TCP *:50050 (LISTEN)
>
> For some reason, the fs and jobtracker are listening on
> localhost.localdomain instead of *. Any ideas on what could be going
> wrong? Also, is there a standard port number to use for the
> fs.default.name, mapred.job.tracker, etc. (I just made them up)?