Member since
08-15-2013
40
Posts
11
Kudos Received
9
Solutions
My Accepted Solutions
Title | Views | Posted |
---|---|---|
7916 | 01-17-2017 10:14 AM | |
2067 | 07-02-2015 05:52 PM | |
2440 | 05-12-2015 10:15 AM | |
2917 | 03-13-2015 11:11 AM | |
4482 | 01-26-2015 07:47 AM |
02-09-2017
11:43 AM
2 Kudos
You can use --ldap_password_cmd="echo <mypassword>" (or something more secure if you so wish). Best, Henry
... View more
01-19-2017
09:39 AM
Filters - as implemented - don't work in the presence of LEFT OUTER JOIN because the filter values are passed from right-to-left. The right side of the join computes the set of filter values, and then the left side is compared against those filter values. But in a LEFT JOIN all left-side values should be output by the join, so no filtering can happen. If you try the same query but with a RIGHT OUTER JOIN, you should see the filters kick back in. Henry
... View more
01-18-2017
12:06 PM
1 Kudo
The profile has the information. Look for something like: Filter 0 (1.00 MB): - Files processed: 76 (76) - Files rejected: 76 (76) - Files total: 76 (76) - RowGroups processed: 0 (0) - RowGroups rejected: 0 (0) - RowGroups total: 0 (0) - Rows processed: 0 (0) - Rows rejected: 0 (0) - Rows total: 0 (0) - Splits processed: 0 (0) - Splits rejected: 0 (0) - Splits total: 0 (0) We don't record per-partition filtering information - but instead we eliminate entire files, so that's captured in the 'Files' statistics below. The 'total' is all files considered for scanning. The 'processed' value is the total number that were submitted to that filter - if the filter arrives late, some files could be read but not considered for filtering. The 'rejected' value is the number of files that were eliminated by the filter. High 'rejected' numbers means an effective filter! HTH, Henry
... View more
01-17-2017
10:14 AM
Can you provide the whole of your 'explain 1'? I would expect to see runtime filtering enabled in the plan. The number of partitions scanned ('partitions=3/3') is correct. Runtime filtering happens at, well, runtime, so the planner doesn't know how many partitions it's going to skip. In your second example, the planner can figure this out ahead of time. Henry
... View more
04-07-2016
02:22 PM
You will need to tell your client (is it shell, JDBC or something else?) to use SSL. Have you read http://www.cloudera.com/documentation/archive/impala/2-x/2-0-x/topics/impala_ssl.html? The instructions for the shell are at the bottom of that pa
... View more
04-06-2016
10:14 PM
This usually means that SSL is configured badly - is your load balancer and / or client configured to make an SSL connection to Impala? The repeated connections are a mystery - they're coming from the client RPC port. So maybe something is using your load balancer repeatedly? Or is the LB itself attempting to create a set of connections (maybe to warm up its connection pool)?
... View more
07-24-2015
11:13 AM
Hi - Not currently, I'm afraid. Authentication and data access go over the same connection. To change that, Impala's clients would have to connect over another channel after authentication. The client would have to receive during authentication some shared secret token allowing it to prove to the server that it's ok to access data. We don't currently have any plans to implement this. Best, Henry
... View more
07-02-2015
05:52 PM
None of Impala's supported file formats are able to store data in sorted order on disk. Therefore the ORDER BY clause in the INSERT does not have any effect. The data is written out in a potentially unsorted order regardless. Best, Henry
... View more
06-24-2015
11:02 AM
You should use --ldap_passwords_in_clear_ok *only* if you're comfortable with Impala sending passwords in the clear to the LDAP server, i.e. not by a secure connection. This flag is mainly as an override for secure-by-default configuration. You should not set it unless there is no way to secure the password-carrying connections. Henry
... View more
05-12-2015
10:15 AM
Have you upgraded Cloudera Manager at the same time? The way in which CM retrieves metric data from Impala changed in CDH5.4, and should be using /jsonmetrics?json as the URL.
... View more