Member since
07-17-2019
738
Posts
433
Kudos Received
111
Solutions
My Accepted Solutions
Title | Views | Posted |
---|---|---|
1995 | 08-06-2019 07:09 PM | |
2281 | 07-19-2019 01:57 PM | |
3026 | 02-25-2019 04:47 PM | |
3513 | 10-11-2018 02:47 PM | |
1043 | 09-26-2018 02:49 PM |
11-04-2016
02:43 PM
Can you please clarify the scenario that you're concerned about? It's not clear at all to me what you're looking for.
... View more
11-03-2016
07:34 PM
Sadly, "Checksum failed" could be for numerous different reasons. You can try to set "-Dsun.security.spnego.debug=true" in PHOENIX_QUERYSERVER_OPTS in hbase-env.sh for lots of debug after a restart of PQS. This will require careful reading and observation (based on principals and hostnames) to figure out what went wrong though.
... View more
11-03-2016
02:57 PM
Hey Chris -- using localhost definitely won't work. You must use the FQDN in the URL you supply (which must match the FQDN in the principal which PQS is using. I'm not sure why you would be getting an HTTP/404 though. I'd take a look at the PQS logs to see if anything is there. Would be best to not piggy-back on this issue and ask a new question instead. Feel free to tag me there so I'm sure to see it.
... View more
10-31-2016
04:16 PM
Yes, the length of the row, not the entire key. The row component in the keys that you provided is "row8". Please refer to the HBase documentation if you are unclear on what components make up an HBase Key.
... View more
10-31-2016
03:36 PM
1 Kudo
It is the length of the row component of the Key.
... View more
10-31-2016
03:01 PM
Just the row component of the Key: http://hbase.apache.org/book.html#_row
... View more
10-28-2016
10:13 PM
Your understanding is correct. Phoenix delegates to HBase permissions. Secondary index tables would require their own grants as well, yes. They are just another HBase table. Yes, you have the steps correct regarding the superuser creating the tables and performing grants. The namespace support actually appears to have landed in 4.8, but this was included in HDP-2.5 which was based on Phoenix 4.7 (the feature was backported) https://phoenix.apache.org/namspace_mapping.html
... View more
10-28-2016
05:09 PM
1. No, SYSTEM tables are created for all Phoenix users. 2. When a user makes the first connection to Phoenix (instantiates the JDBC driver) it will check and try to create the SYSTEM tables if they don't already exist. For all but the first connection, this will be a no-op. If you have permissions put in place, you will want to launch sqlline (or some application using Phoenix) which has the permission to create these SYSTEM tables. Then, before having unprivileged users access phoenix, make sure they have read permission on the system tables. 3. You can create an HBase Namespace, grant privileges on that namespace to a user, and then instruct the user to create phoenix tables like "MY_NAMESPACE.MY_TABLE".
... View more
10-26-2016
02:45 PM
Hi @Klaus Lucas. Can you also check the Accumulo Master log around 2016-10-26 14:56:50,314? The FATE operation being performed here is an RPC to the Accumulo Master process. The TApplicationException thrown by the Tracer waiting on the response from that RPC indicates that there was an uncaught exception thrown by the Accumulo Master. Figuring out what happened in the Master will help us understand what wrong wrong and caused the tracer to crash.
... View more
10-25-2016
03:22 PM
VARCHAR's cannot be naturally converted to numbers. That is what is meant by "VARCHAR cannot be coerced to DOUBLE". Use the TO_NUMBER function when inserting data: http://phoenix.apache.org/language/functions.html#to_number
... View more