asterixdb-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ben Tovar <>
Subject Re: Installation debug output
Date Fri, 07 Aug 2015 11:52:01 GMT

Yes, I did a configure. I can run so that the zookeper directory is
writable , but I think the problem is that the ssh-keys do not
authenticate us for AFS, so the shell init scripts cannot be read.


I think that also might be it. Yesterday I tried on a FreeBSD box,
which does not install bash at /bin/bash (or at all by default), and I
saw that /bin/bash was being called directly. On the linux side I am
using zsh, so that might be another possibility.

As a test, if I create a local user, with bash as the shell, then the
basic tutorial work as expected. This is really nice, but I would
really like to try the many workers configuration :).

By the way, on FreeBSD I got farther. I got an unusable instance with
a message that the cluster controller was not running. I imagine this
is not one of the main targeted platforms, but I can test there if you


On Fri, Aug 7, 2015 at 2:18 AM, Ian Maxon <> wrote:
> Also, if this is the last stable version (0.8.6), that error can
> happen if your shell happens not to be bash. This is mitigated in the
> latest snapshots (but I have still had issues under fish).
> -Ian
> On Thu, Aug 6, 2015 at 10:44 PM, Pouria Pirzadeh
> <> wrote:
>> Ben,
>> Did run 'managix configure' command at the beginning ? And if yes, can you
>> make sure 'managix-conf.xml' file under $MANAGIX_HOME/conf directory
>> contains the correct java_home and a writable zookeeper directory ?
>> w.r.t the password-less SSH, basically the username that wants to launch
>> asterixdb processes (whoever runs managix) should be able to do SSH without
>> prompting for any password to all the machines that you have included in the
>> local.xml file (including localhost or , if it is there).
>> Pouria
>> On Thu, Aug 6, 2015 at 4:53 AM, Ben Tovar <> wrote:
>>> Hi,
>>> I am trying the single machine installation example, and the
>>> managix create -n my_asterix -c $MANAGIX_HOME/clusters/local/local.xml
>>> fails telling me that the zookeeper service was unable to start.
>>> java_home points to the correct location, and
>>> managix validate
>>> does not report any errors. Is there a log I can look at (or enable)
>>> to see why create is failing?
>>> (That said, my suspicion is that ssh-keys do not play nicely with
>>> kerberos authentication, which our site uses. Does anybody have
>>> experience with that? Are there other authentication non-ssh methods I
>>> can try?)
>>> Best wishes!,
>>> Ben

View raw message