Somewhere along the line, while trying out different Cassandra/DSE implementations and versions, I messed up my config files.
Accordingly, when I run
./dse nodetool status
it returns:
/home/ubuntu/lib/dse-6.8.0/resources/cassandra/bin/cassandra.in.sh: line 75: /etc/dse/cassandra/jvm-dependent.sh: No such file or directory
Is there a way to get this config file back? Or better, just to totally refresh my DSE 6.8 install so it puts all the default files back?
It isn't an installation issue. The problem is that this command is invalid:
$ dse nodetool status
The dse tool is for running commands like starting/stopping DSE and launching clients (such as a Spark shell) to connect to a node running DataStax Enterprise.
The nodetool utility is for retrieving information or performing operations on Cassandra. The dsetool utility provides a similar functionality but retrieves information about DSE components or running DSE-specific operations.
To get information about a Cassandra cluster, run:
$ nodetool status
For example:
$ nodetool status Datacenter: Cassandra ===================== Status=Up/Down |/ State=Normal/Leaving/Joining/Moving/Stopped -- Address Load Owns (effective) Host ID Token Rack UN 10.101.34.115 277.27 MiB 100.0% 38d53689-8039-470c-91c2-4ede83066826 3835820215023312942 rack1
To get additional information about a DSE cluster such as workload type and node health, run:
$ dsetool status
For example:
$ dsetool status DC: Cassandra Workload: Cassandra Graph: no ====================================================== Status=Up/Down |/ State=Normal/Leaving/Joining/Moving -- Address Load Effective-Ownership Token Rack Health [0,1] UN 10.101.34.115 277.27 MiB 100.00% 3835820215023312942 rack1 1.00
For more information, see DataStax Enterprise Tools. Cheers!
Oh yeah, I always forget that it's not officially supported to do it that way. It is strange that it still runs though. Regardless, I get the same result if I do it the correct way:
I can change the question though if it's helpful.
Thanks
OK, so here's something really curious. You're running the commands in bin/ as if you've got a tarball installation. If this is true, then the jvm-dependent.sh script should be in the conf/ directory.
But the cassandra.in.sh script that gets called is looking for jvm-dependent.sh in /etc/dse/cassandra which indicates that it's a packaged installation.
Can you tell me how you installed DSE 6.8.0? Also, did you remove the previous package installation(s) before installing DSE 6.8.0?
After looking closely, it indicates that the installations are scrambled on the node. Cheers!
Yeah so I have ran DSE on docker, using the debian package installation, and from the tarball all on this Ubuntu server before, as I was playing around with things. Currently I am using the tarball as you mentioned.
I have since removed the docker images (which shouldn't affect things anyways I think) and the package installation (though probably not correctly, hence the errors). It's been too long since I've done it to remember how I removed the package installation as well.
Given that, is there a way to fix it? Or is it beyond repair at this point?
Thanks again though, even diagnosing that the `cassandra.in.sh` script is looking in the wrong place is helpful in itself
6 People are following this question.
