hadoop-hdfs-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Daniel Watrous <dwmaill...@gmail.com>
Subject Re: Problem running example (wrong IP address)
Date Mon, 28 Sep 2015 14:07:12 GMT
Vinay,

There is no gateway to 51.*. These are IP addresses that I set in my
Vagrantfile for virtualbox as part of a private network:
master.vm.network "private_network", ip: 192.168.51.4

This allows me to spin up all the hosts for my cluster automatically and
know that they always have the same IP addresses.

>From hadoop-data1 (192.168.52.4) I have unrestricted access to
hadoop-master (192.168.51.4)

hadoop@hadoop-data1:~$ ifconfig
eth1      Link encap:Ethernet  HWaddr 08:00:27:b9:55:25
          inet addr:192.168.52.4  Bcast:192.168.52.255  Mask:255.255.255.0
hadoop@hadoop-data1:~$ ping hadoop-master
PING hadoop-master (192.168.51.4) 56(84) bytes of data.
64 bytes from hadoop-master (192.168.51.4): icmp_seq=1 ttl=63 time=3.13 ms
64 bytes from hadoop-master (192.168.51.4): icmp_seq=2 ttl=63 time=2.72 ms

I'm not sure I understand exactly what you're asking for, but from the
master I can run this

vagrant@hadoop-master:~$ sudo netstat -tnulp | grep 54310
tcp        0      0 0.0.0.0:54310           0.0.0.0:*               LISTEN
     22944/java

I understand what you're saying about a gateway often existing at that
address for a subnet. I'm not familiar enough with Vagrant to answer this
right now, but I will put in a question there.

I can also change the other two IP addresses to be on the same 51. subnet.
I may try that next.



On Mon, Sep 28, 2015 at 8:33 AM, Vinayakumar B <vinayakumarb@apache.org>
wrote:

> 192.168.51.1 might be gateway to 51.* subnet right?
>
> Can you verify whether connections from outside 51 subnet, to 51.4 machine
> using other subnet IP as remote IP. ?
>
> You can create any connection, may not be namenode-datanode.
>
> for ex: Connection from 192.168.52.4 dn to 192.168.51.4 namenode should
> result in following, when checked using netstat command in namenode
> machine. "netstat -tnulp | grep <NN_RPC_PORT>"
>
> Output should be something like below
>
> tcp        0      0   192.168.51.4:54310        192.168.52.4:32567
>     LISTEN      -
>
>
> If the Foreign Ip is listing as 192.168.51.1 instead of 192.168.52.4, then
> the gateway, is not passing original client IP forward, its re-creating
> connections with its own IP. in such case problem will be with the gateway.
>
> Its just a guess, reality could be different.
>
> please check and let me know.
>
> -Vinay
>
> On Mon, Sep 28, 2015 at 6:45 PM, Daniel Watrous <dwmaillist@gmail.com>
> wrote:
>
>> Thanks to Namikaze pointing out that I should have sent the namenode log
>> as a pastbin
>>
>> http://pastebin.com/u33bBbgu
>>
>>
>> On Mon, Sep 28, 2015 at 8:02 AM, Daniel Watrous <dwmaillist@gmail.com>
>> wrote:
>>
>>> I have posted the namenode logs here:
>>> https://gist.github.com/dwatrous/dafaa7695698f36a5d93
>>>
>>> Thanks for all the help.
>>>
>>> On Sun, Sep 27, 2015 at 10:28 AM, Brahma Reddy Battula <
>>> brahmareddy.battula@hotmail.com> wrote:
>>>
>>>> Thanks for sharing the logs.
>>>>
>>>> Problem is interesting..can you please post namenode logs and dual IP
>>>> configurations(thinking problem with gateway while sending requests from
>>>> 52.1 segment to 51.1 segment..)
>>>>
>>>> Thanks And Regards
>>>> Brahma Reddy Battula
>>>>
>>>>
>>>> ------------------------------
>>>> Date: Fri, 25 Sep 2015 12:19:00 -0500
>>>>
>>>> Subject: Re: Problem running example (wrong IP address)
>>>> From: dwmaillist@gmail.com
>>>> To: user@hadoop.apache.org
>>>>
>>>> hadoop-master http://pastebin.com/yVF8vCYS
>>>> hadoop-data1 http://pastebin.com/xMEdf01e
>>>> hadoop-data2 http://pastebin.com/prqd02eZ
>>>>
>>>>
>>>>
>>>> On Fri, Sep 25, 2015 at 11:53 AM, Brahma Reddy Battula <
>>>> brahmareddy.battula@hotmail.com> wrote:
>>>>
>>>> sorry,I am not able to access the logs, could please post in paste bin
>>>> or attach the 192.168.51.6( as your query is why different IP) DN logs
>>>> and namenode logs here..?
>>>>
>>>>
>>>>
>>>>
>>>> Thanks And Regards
>>>> Brahma Reddy Battula
>>>>
>>>>
>>>> ------------------------------
>>>> Date: Fri, 25 Sep 2015 11:16:55 -0500
>>>> Subject: Re: Problem running example (wrong IP address)
>>>> From: dwmaillist@gmail.com
>>>> To: user@hadoop.apache.org
>>>>
>>>>
>>>> Brahma,
>>>>
>>>> Thanks for the reply. I'll keep this conversation here in the user
>>>> list. The /etc/hosts file is identical on all three nodes
>>>>
>>>> hadoop@hadoop-data1:~$ cat /etc/hosts
>>>> 127.0.0.1 localhost
>>>> 192.168.51.4 hadoop-master
>>>> 192.168.52.4 hadoop-data1
>>>> 192.168.52.6 hadoop-data2
>>>>
>>>> hadoop@hadoop-data2:~$ cat /etc/hosts
>>>> 127.0.0.1 localhost
>>>> 192.168.51.4 hadoop-master
>>>> 192.168.52.4 hadoop-data1
>>>> 192.168.52.6 hadoop-data2
>>>>
>>>> hadoop@hadoop-master:~$ cat /etc/hosts
>>>> 127.0.0.1 localhost
>>>> 192.168.51.4 hadoop-master
>>>> 192.168.52.4 hadoop-data1
>>>> 192.168.52.6 hadoop-data2
>>>>
>>>> Here are the startup logs for all three nodes:
>>>> https://gist.github.com/dwatrous/7241bb804a9be8f9303f
>>>> https://gist.github.com/dwatrous/bcd85cda23d6eca3a68b
>>>> https://gist.github.com/dwatrous/922c4f773aded0137fa3
>>>>
>>>> Thanks for your help.
>>>>
>>>>
>>>> On Fri, Sep 25, 2015 at 10:33 AM, Brahma Reddy Battula <
>>>> brahmareddy.battula@huawei.com> wrote:
>>>>
>>>> Seems DN started in three machines and failed in
>>>> hadoop-data1(192.168.52.4)..
>>>>
>>>>
>>>> 192.168.51.6 : giving IP as 192.168.51.1 <http://192.168.51.1:50010>...can
>>>> you please check /etc/hosts file of 192.168.51.6 (might be 192.168.51.1
>>>> <http://192.168.51.1:50010> is configured in /etc/hosts)
>>>>
>>>> 192.168.52.4 : datanode startup might be failed ( you can check this
>>>> node logs)
>>>>
>>>> 192.168.51.4 :  <http://192.168.51.4:50010> Datanode starup is
>>>> success..which is in master node..
>>>>
>>>>
>>>>
>>>> Thanks & Regards
>>>>  Brahma Reddy Battula
>>>>
>>>>
>>>>
>>>> ------------------------------
>>>> *From:* Daniel Watrous [dwmaillist@gmail.com]
>>>> *Sent:* Friday, September 25, 2015 8:41 PM
>>>> *To:* user@hadoop.apache.org
>>>> *Subject:* Re: Problem running example (wrong IP address)
>>>>
>>>> I'm still stuck on this and posted it to stackoverflow:
>>>>
>>>> http://stackoverflow.com/questions/32785256/hadoop-datanode-binds-wrong-ip-address
>>>>
>>>> Thanks,
>>>> Daniel
>>>>
>>>> On Fri, Sep 25, 2015 at 8:28 AM, Daniel Watrous <dwmaillist@gmail.com>
>>>> wrote:
>>>>
>>>> I could really use some help here. As you can see from the output
>>>> below, the two attached datanodes are identified with a non-existent IP
>>>> address. Can someone tell me how that gets selected or how to explicitly
>>>> set it. Also, why are both datanodes shown under the same name/IP?
>>>>
>>>> hadoop@hadoop-master:~$ hdfs dfsadmin -report
>>>> Configured Capacity: 84482326528 (78.68 GB)
>>>> Present Capacity: 75745546240 (70.54 GB)
>>>> DFS Remaining: 75744862208 (70.54 GB)
>>>> DFS Used: 684032 (668 KB)
>>>> DFS Used%: 0.00%
>>>> Under replicated blocks: 0
>>>> Blocks with corrupt replicas: 0
>>>> Missing blocks: 0
>>>> Missing blocks (with replication factor 1): 0
>>>>
>>>> -------------------------------------------------
>>>> Live datanodes (2):
>>>>
>>>> Name: 192.168.51.1:50010 (192.168.51.1)
>>>> Hostname: hadoop-data1
>>>> Decommission Status : Normal
>>>> Configured Capacity: 42241163264 (39.34 GB)
>>>> DFS Used: 303104 (296 KB)
>>>> Non DFS Used: 4302479360 (4.01 GB)
>>>> DFS Remaining: 37938380800 (35.33 GB)
>>>> DFS Used%: 0.00%
>>>> DFS Remaining%: 89.81%
>>>> Configured Cache Capacity: 0 (0 B)
>>>> Cache Used: 0 (0 B)
>>>> Cache Remaining: 0 (0 B)
>>>> Cache Used%: 100.00%
>>>> Cache Remaining%: 0.00%
>>>> Xceivers: 1
>>>> Last contact: Fri Sep 25 13:25:37 UTC 2015
>>>>
>>>>
>>>> Name: 192.168.51.4:50010 (hadoop-master)
>>>> Hostname: hadoop-master
>>>> Decommission Status : Normal
>>>> Configured Capacity: 42241163264 (39.34 GB)
>>>> DFS Used: 380928 (372 KB)
>>>> Non DFS Used: 4434300928 (4.13 GB)
>>>> DFS Remaining: 37806481408 (35.21 GB)
>>>> DFS Used%: 0.00%
>>>> DFS Remaining%: 89.50%
>>>> Configured Cache Capacity: 0 (0 B)
>>>> Cache Used: 0 (0 B)
>>>> Cache Remaining: 0 (0 B)
>>>> Cache Used%: 100.00%
>>>> Cache Remaining%: 0.00%
>>>> Xceivers: 1
>>>> Last contact: Fri Sep 25 13:25:38 UTC 2015
>>>>
>>>>
>>>>
>>>> On Thu, Sep 24, 2015 at 5:05 PM, Daniel Watrous <dwmaillist@gmail.com>
>>>> wrote:
>>>>
>>>> The IP address is clearly wrong, but I'm not sure how it gets set. Can
>>>> someone tell me how to configure it to choose a valid IP address?
>>>>
>>>> On Thu, Sep 24, 2015 at 3:26 PM, Daniel Watrous <dwmaillist@gmail.com>
>>>> wrote:
>>>>
>>>> I just noticed that both datanodes appear to have chosen that IP
>>>> address and bound that port for HDFS communication.
>>>>
>>>> http://screencast.com/t/OQNbrWFF
>>>>
>>>> Any idea why this would be? Is there some way to specify which
>>>> IP/hostname should be used for that?
>>>>
>>>> On Thu, Sep 24, 2015 at 3:11 PM, Daniel Watrous <dwmaillist@gmail.com>
>>>> wrote:
>>>>
>>>> When I try to run a map reduce example, I get the following error:
>>>>
>>>> hadoop@hadoop-master:~$ hadoop jar
>>>> /usr/local/hadoop/share/hadoop/mapreduce/hadoop-mapreduce-examples-2.7.1.jar
>>>> pi 10 30
>>>> Number of Maps  = 10
>>>> Samples per Map = 30
>>>> 15/09/24 20:04:28 INFO hdfs.DFSClient: Exception in
>>>> createBlockOutputStream
>>>> java.io.IOException: Got error, status message , ack with firstBadLink
>>>> as 192.168.51.1:50010
>>>>         at
>>>> org.apache.hadoop.hdfs.protocol.datatransfer.DataTransferProtoUtil.checkBlockOpStatus(DataTransferProtoUtil.java:140)
>>>>         at
>>>> org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.createBlockOutputStream(DFSOutputStream.java:1334)
>>>>         at
>>>> org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.nextBlockOutputStream(DFSOutputStream.java:1237)
>>>>         at
>>>> org.apache.hadoop.hdfs.DFSOutputStream$DataStreamer.run(DFSOutputStream.java:449)
>>>> 15/09/24 20:04:28 INFO hdfs.DFSClient: Abandoning
>>>> BP-852923283-127.0.1.1-1443119668806:blk_1073741825_1001
>>>> 15/09/24 20:04:28 INFO hdfs.DFSClient: Excluding datanode
>>>> DatanodeInfoWithStorage[192.168.51.1:50010
>>>> ,DS-45f6e06d-752e-41e8-ac25-ca88bce80d00,DISK]
>>>> 15/09/24 20:04:28 WARN hdfs.DFSClient: Slow waitForAckedSeqno took
>>>> 65357ms (threshold=30000ms)
>>>> Wrote input for Map #0
>>>>
>>>> I'm not sure why it's trying to access 192.168.51.1:50010, which isn't
>>>> even a valid IP address in my setup.
>>>>
>>>> Daniel
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>
>

Mime
View raw message