Vertica 7: 'NoneType' object is not subscriptable on create database. Ubuntu 12.04

Hi Joe, You've reached the Vertica community forums. If you're running into trouble, could you describe your situation in more detail? What commands you are running to get here, anything unusual about your environment, etc. Adam

Vertica Community Edition 7; Ubuntu 12.04 64-bit running under VMware Workstation 10.
Vertica was not installed on this machine before.
The above error is reported when trying to create (the first) database on the above configuration.
The directories and files are created, but, unusable due to the above error, I presume.
I have Vertica 6 running successfully on another box.

Thank you for your response, John
PS: ex HP

PPS: you can write directly to johndapps at gmail dot com

Thanks John; could you post the commands that you're running to get this message?

I suspect that this is happening because of some argument to some script that's not in a format that we're expecting.  (I've installed Vertica 7 on Ubuntu without any trouble.)

If you're running under VMware Workstation, we actually provide a VMware image for you; you might find it easier to use.  (On the Downloads page; you have to scroll down a bit.)  It's CentOS-based, though; not Ubuntu.

If you'd like support via e-mail, you can file a support case at my.vertica.com.  (We keep Community discussions online so that others can jump in and can see the resolutions.)

I did try the VMware image. However, it is very old compared to what I am running and the display size in CentOS is setup for people who ware magnifying glasses...
I'll persevere with VMware Workstartion 10 and Ubuntu for now, but thank you for pointing out that there is a VMware image to play with.
Next stop might be AWS, just to give it a go and see what else I can break

Hm...  And how did you invoke install_vertica initially?

I'm not sure how you get into a state where you have nodes but adminTools can't see them.  But you could poke around in /opt/vertica/config/ .

Well, your thought about checking the config file was a good one. What had happened was that I had specified the node name (Ubuntu) rather than localhost.
Ubuntu was assigned an address of 127.0.1.1 versus localhost with 127.0.0.1
The above values were picked up by the adminTools script...
I changed the hosts to 127.0.0.1 and now everything appears to work.
In addition to the above changes, I also changed the default_base which had been set to a value different from the directory allocated to the nodes.

Not sure what caused the above mess, but, moral of the story appears to be: do not use a host name unless you are sure it is the one which Vertica will use too!

Cheers, and thank you for the help, John

Ah, glad you were able to sort this out!

I am getting the same problem when I try to replace a defective server (10.0.28.87) in the cluster with 10.0.28.89 That server was not responding to ssh requests and has been taken offline. I'm using Vertica7 on Linux. Here's the output: Unable to SSH as dbadmin to 10.0.28.87: UNEXPECTED EOF: Could not login with ssh. ssh output: 'ssh: connect to host 10.0.28.87 port 22: No route to host\r\r\n' 10.0.28.86 OK [vertica][(7, 0, 0)][001][x86_64] 10.0.28.88 OK [vertica][(7, 0, 0)][001][x86_64] 10.0.28.89 OK [vertica][(7, 0, 0)][001][x86_64] adminTools Last Chance Error Handler running... raised error: error message: 'NoneType' object is not subscriptable trace file: /opt/vertica/log/adminTools-dbadmin.errors REPORT THIS INFORMATION TO TECHNICAL SUPPORT AND INCLUDE CONTENTS OF THE TRACE FILE IN YOUR REPORT The /opt/vertica/log/adminTools-dbadmin.errors file has these lines: -------------------------------- handler invoked: 2014-01-12 22:17:06 raised error: Traceback (most recent call last): File "./admintools", line 374, in ac.runApp() File "/opt/vertica/oss/python/lib/python2.7/site-packages/vertica/ui/adminCtrl.py", line 3409, in runApp self.__ui.runFromNavigation("main") File "/opt/vertica/oss/python/lib/python2.7/site-packages/vertica/ui/uiMgr.py", line 165, in a_wrapper_accepting_args fn(self, *args, **kwargs) File "/opt/vertica/oss/python/lib/python2.7/site-packages/vertica/ui/uiMgr.py", line 450, in runFromNavigation self.__navigator.navigate(uimanager=self, startingPoint=runFrom) File "/opt/vertica/oss/python/lib/python2.7/site-packages/vertica/ui/Navigator.py", line 315, in navigate ((self.__menus[ nextMenu ][ self.NAV_MACTIONS ])[ offset ])() File "/opt/vertica/oss/python/lib/python2.7/site-packages/vertica/ui/adminCtrl.py", line 3320, in clusterReplaceNode (code, msg) = self.__exec.replaceNodeDefinition(dbname, hosts, replacements, upassword) File "/opt/vertica/oss/python/lib/python2.7/site-packages/vertica/engine/adminExec.py", line 2494, in replaceNodeDefinition """ % ( site_name, new_host, netprofile[new_host][0], netprofile[new_host][1] ) TypeError: 'NoneType' object is not subscriptable

Hi Brian!

Please review this thread: https://community.vertica.com/vertica/topics/single_node_cluster_change_node_ip_address-14fuee
You need to change IP addresses in catalog too.


Regards Daniel

Thanks Daniel, this is useful information. I didn't know we could restore a backup to a cluster with different ip addresses. 

Is this what you recommend for a down server?
1) use 1 spare server. Change the ip for the down server to the new server, on all servers in the cluster in the following files.  Also copy these to the replacement server.

• /opt/vertica/config/admintools.conf
• /opt/vertica/config/vspread.conf
• /etc/sysconfig/spreadd

2) Restart the cluster

I worry that step 1 will leave the cluster in a state where it
does not realize that data is lost and needs recovery on
one of the servers.
Will auto-recovery start in this case?
Or should I follow with

3) Run the admintools - replace node to replace the
just-replaced server with another.

Thanks for your time,
Brian

Hm...no...

I read one more time and understood that I didn't get you previously and
version 7 probably has an easier way.

Common actions its:

1. Low your K-Safety
https://my.vertica.com/docs/7.0.x/HTML/index.htm#Authoring/AdministratorsGuide/ManageNodes/LoweringT...

2. Remove host from cluster
https://my.vertica.com/docs/7.0.x/HTML/index.htm#Authoring/AdministratorsGuide/RemovingHostsFromAClu...

3. Remove host from database
https://my.vertica.com/docs/7.0.x/HTML/index.htm#Authoring/AdministratorsGuide/RemovingNodesFromADat...

4. Change IP in configuration
https://my.vertica.com/docs/7.0.x/HTML/index.htm#Authoring/AdministratorsGuide/ManageNodes/ChangingT...

im apologize for confusing you, but now you know also other way (via catalog editor)


Edit IP in configurations and try FORCE recovery:

daniel@synapse ~ $ adminTools -t restart_node -h
Usage: restart_node [options]

Options:
  -h, --help            show this help message and exit
  -s NODES, --hosts=NODES
                        comma-separated list of hosts to be restarted
  -d DB, --database=DB  Name of database whose node is to be restarted
  -p DBPASSWORD, --password=DBPASSWORD
                        Database password in single quotes
  -i, --noprompts       do not stop and wait for user input(default false)
-F, --force           force the node to start and auto recover if necessary
  --compat21            Use Vertica 2.1 method using node names instead of
                        hostnames

Thanks Daniel,

Removing host often requires the host be up.
Anyway, since I have an SRE team, they are trying to bring
up the down node.
If they fail, they will give me a fresh server with the same IP.
So in the end things will be fine.
I was just hoping to replace the server as an interim solution.

Glad someone is helping on these forums,
Brian