[
https://issues.apache.org/jira/browse/TRAFODION-1052?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14717283#comment-14717283
]
Suresh Subbiah commented on TRAFODION-1052:
-------------------------------------------
Judy Zhao (hongxia-zhao) on 2015-03-17
Changed in trafodion:
status: New → In Progress
Judy Zhao (hongxia-zhao) wrote on 2015-03-17: #1
Root cause:
For mxosrvr, it defined following domain suffix as usual top level domain names
in host_addr_get funciotn. The hostname <n007.cm.cluster> isn’t listed here.
char *topDomain[] = {".com", ".net", ".org", ".edu", ".gov", ".biz", ".mil",
".name"};
With the last fixing for [Bug 1415936] , the function host_addr_get will be
called every time. So the issue was exposed.
How to fix:
According to Matt:
1.Function host_addr_get() was a bug that bit us long ago and it should not
ever be called. Reason is we cannot ever possibly account for all the DNS
suffixes in the world.
2.TCPADD is always passed to mxosrvr with IP and hostname. We did that so
mxosrvr wouldn’t have to discover anything.
So the fixing is :
1.Removed host_addr_get() .
2.To get hostname from parameter '-RZ" .
3.The parameter hostname or TCPID are requireds, if mxosrvr doesn't find the
parameters it will exit.
Chirag Bhalgami (chirag-bhalgami) on 2015-03-19
tags: added: installer
Judy Zhao (hongxia-zhao) on 2015-03-20
Changed in trafodion:
status: In Progress → Fix Committed
Marvin Anderson (marvin-anderson) on 2015-03-31
tags: removed: installer
> LP Bug: 1430900 - A network component error host_addr_get failed has occurred
> -----------------------------------------------------------------------------
>
> Key: TRAFODION-1052
> URL: https://issues.apache.org/jira/browse/TRAFODION-1052
> Project: Apache Trafodion
> Issue Type: Bug
> Reporter: Chirag Bhalgami
> Assignee: JudyZhao
> Priority: Blocker
>
> After installing Trafodion, DCS doesn't start mxosrvr.
> DCS log doesn't have any ERROR but Trafodion log has following error.
> --------------
> [trafodion@n007 ~]$ cat
> /opt/trafodion/trafodion-20150310_0830/logs/master_exec_0_21748.log
> ERROR, MXOSRVR, Node Number: 0, CPU: 0, PIN:12499, Process Name:$Z000A74 , ,
> ,A network component error host_addr_get failed has occurred. CEE Error Text:
> <5>.
> [trafodion@n007 ~]$ hostname
> n007
> [trafodion@n007 ~]$ hostname -f
> n007.cm.cluster
> [trafodion@n007 ~]$ ping n007.cm.cluster
> PING n007.cm.cluster (<global_ip_address>) 56(84) bytes of data.
> 64 bytes from n007.cm.cluster (<global_ip_address>): icmp_seq=1 ttl=64
> time=0.024 ms
> 64 bytes from n007.cm.cluster (<global_ip_address>): icmp_seq=2 ttl=64
> time=0.195 ms
> --------------
> The cluster has CDH 5.3.0
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
