asterixdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ian Maxon <>
Subject Re: Unable to create an instance
Date Thu, 02 Jul 2015 01:01:54 GMT
Ah, a fellow fish user :) I have had the same issues, for the same reason then.
Basically, the root of the cause is that Managix assume(d) wrongly, in
many situations, that bash was the default shell. For example the
script I pointed at, you probably noticed it lacked a shebang- hence
the default interpreter is used, which is not exactly predictable.

I did submit a patch for this exact issue (
, and it got merged, but it doesn't seem to solve the problem
completely. There are probably other instances of assuming bash is the
default shell, in the Java part of Managix rather than the scripted
part. One thing I suppose that could be tried, is to use the latest
snapshot version (or just apply the changes from that manually).
Whenever I must use managix though, I typically am forced to switch to
bash for a while.

In lieu of changing shells (which is painful and cumbersome...), it
might be easier to try just running AsterixDB locally, without
managix, since it's not necessary in this case: . This
is a packaging that is simpler to run, just for demo purposes. All you
have to do is extract the archive, and then cd into the bin/ folder
and run It will start up a CC and 2 NCs. To exit just
Ctrl+C and it will shut down all 3 started processes.

- Ian

On Wed, Jul 1, 2015 at 4:49 PM, Hanifi GUNES <> wrote:
> Thanks Ian. Right. I am on fish and not bash/sh. Running my pre-installed
> ZK manually and changing the ZK client port, I was able to by-pass the
> previous failure. However, I am hitting to another one complaining about
> node state this time. Log dir at $MANAGIX_HOME/clusters/local/local.xml
> does not seem to exist.
> Any pointers?
> *> managix start -n mydbDEBUG: ZooKeeper running at
> <>INFO: Name:mydbCreated:Wed Jul 01 16:42:31 PDT
> 2015Web-Url: <>State:UNUSABLE*
> *WARNING!:Cluster Controller not running at masterNode Controller not
> running at the following nodes127.0.0.1*
> 2015-07-01 15:50 GMT-07:00 Ian Maxon <>:
>> Hi Hanafi,
>> Sorry to hear you're running into troubles getting started up. That
>> error is unfortunately less specific than it might suggest. One of the
>> most common reasons this happens is if you are using a shell besides
>> bash or sh. There's a script in $MANAGIX_HOME/.installer/zookeeper/bin
>> that is actually trying to start Zookeeper. One thing I usually try,
>> when this step is being especially tempermental, is to just run that
>> script with reasonable parameters by hand. After Zookeeper gets
>> started up, everything else will proceed normally.
>> There's also other ways to start AsterixDB just for demo purposes (as
>> a local process, in a Docker container, etc...), so if the above
>> doesn't work, I can point you at some of those alternatives.
>> Thanks,
>> -Ian
>> On Wed, Jul 1, 2015 at 3:24 PM, Hanifi GUNES <> wrote:
>> > I am following the instructions given at [1] to get Asterix up and
>> running
>> > on my Mac. For some reason, I cannot get managix to create an instance. I
>> > already set the log level down to ALL and made sure java_home is exported
>> > to shell as well as to conf/managix-conf.xml
>> >
>> > Can someone point me to right direction on how to debug this?
>> >
>> >
>> > *> managix validate*
>> > *INFO: Environment [OK]**INFO: Managix Configuration [OK]*
>> >
>> >
>> > *> managix create -n mydb -c $MANAGIX_HOME/clusters/local/local.xml*
>> > *DEBUG: Starting ZooKeeper at <>*
>> > *ERROR: Unable to start Zookeeper Service. This could be because of the
>> > following reasons.*
>> > *1) Managix is incorrectly configured. Please run managix validate to
>> run a
>> > validation test and correct the errors reported.**2) If validation in (1)
>> > is successful, ensure that java_home parameter is set correctly in
>> Managix
>> > configuration (conf/managix-conf.xml)*
>> >
>> > Thanks
>> > -Hanifi
>> >
>> >
>> > 1:

View raw message