1

I'm building DataStax Enterprise backup / restore utilities.

nodetool snapshot works fine.

the snapshot restore process is overly complex.

I am trying to get https://github.com/gianlucaborello/cassandradump to work. The cassandradump utility is working fine, although I get the error:

UnicodeDecodeError: 'utf8' codec can't decode byte 0xff in position 42: invalid start byte

when it gets to the backup of the system.local table.

So, my question is: how do I set the encoding for my cluster, or my keyspace, to UTF-8? To test the cassandradump utility, I am using the command:

python cassandradump.py --keyspace system --export-file testdump.cql

I fear the answer may be that the encoding of the system keyspace is not my concern, but shouldn't that be configurable?

I have also seen this post: UnicodeDecodeError: 'utf8' codec can't decode bytes in position 3-6: invalid data

But, it doesn't seem applicable to the cassandradump utility. It looks like the cassandradump utility has tried to allow for encoding as best it can.

Community
  • 1
  • 1
fcnorman
  • 1,154
  • 9
  • 19
  • 1) Not sure why you're backing the system keyspace, messing with the system keyspace in general is not a good idea unless you know what you're doing 2) If you're using DSE, the supported approach is the OpsCenter backup service which is built in. – phact Feb 03 '16 at 19:26

0 Answers0