zookeeper-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Patrick Hunt <ph...@apache.org>
Subject Re: Killing a zookeeper server
Date Mon, 25 Jan 2010 21:44:44 GMT
JD, there's something _very_ unusual in your setup. Are you running 
"official" released ZooKeeper code or something else?

Either there is a misconfiguration on the other servers (the configs for 
the other servers is exactly the same as 222 right?), or perhaps some 
patches to ZK codebase that went awry?

See the attached file "zk_ports.txt". This is a summary of the netstat 
-a that you sent. Notice in particular that UDP sockets are open for 
port 2888! This should not happen in the default ZK configuration case.

By default we only use tcp connections between servers (quorum & 
election). There is a "electionAlg" option that allows users to turn off 
the TCP based fast leader election and go with a UDP based, but I don't 
see that in the config you provided for 222. (as I said, assuming you 
are not setting this option on the other servers either, correct?).


Mahadev and I do remember that there was a bug in the 3.2 branch prior 
to 3.2 ever being released that caused us to use non-FLE (so UDP based) 
election by default, however we fixed that before 3.2.0 ever shipped (it 
was a bug in our config processing code) and it was never exposed in an 
official release. Perhaps you have picked up some code prior to that?

Patrick

Jean-Daniel Cryans wrote:
>> According to the log for 222 it can't open a connection to the election port
>> (3888) for any of the other servers. This seems very unusual. Can you verify
>> that ther's connectivity on that port btw 222 and all the other servers?
> 
> jdcryans@sv4borg222:~$ telnet sv4borg224 3888
> Trying 10.10.20.224...
> telnet: Unable to connect to remote host: Connection refused
> jdcryans@sv4borg222:~$ telnet sv4borg224 2888
> Trying 10.10.20.224...
> Connected to sv4borg224.
> Escape character is '^]'.
> 
>> Also, can you re-run the netstat with -a option? We can see the listen
>> sockets that way (omitted by netstat by default). It would be great if you
>> could send the netstat for all 5 servers.
> 
> I updated the tar.gz with the 5 netstat -anp
> 
> Thx!
> 
> J-D
> 
>> Thanks,
>>
>> Patrick
>>
>> Jean-Daniel Cryans wrote:
>>> Everything is here
>>> http://people.apache.org/~jdcryans/zk_election_bug.tar.gz
>>>
>>> The server we are trying to start is sv4borg222 (myid is 2) and we
>>> started it around 10:03:21
>>>
>>> Thx!
>>>
>>> J-D
>>>

Mime
View raw message