I just started using the driver org.apache.hive.jdbc.HiveDriver (Version
1.2.1 for spark2) with a Spark Thrift Server (STS) (reference here)
java.sql.ResultSet defines the method absolute() (JavaDoc here)
but HiveBaseResultSet seems to have chosen not to implement the method (source code here)
So now my application (built on top of SmartGWT) was doing a simple operation and I got the following error message:
=== 2017-05-13 18:06:16,980 [3-47] WARN RequestContext - dsRequest.execute() failed:
java.sql.SQLException: Method not supported
at org.apache.hive.jdbc.HiveBaseResultSet.absolute(HiveBaseResultSet.java:70)
at org.apache.commons.dbcp.DelegatingResultSet.absolute(DelegatingResultSet.java:373)
at com.isomorphic.sql.SQLDataSource.executeWindowedSelect(SQLDataSource.java:2970)
at com.isomorphic.sql.SQLDataSource.SQLExecute(SQLDataSource.java:2024)
What is the reason that the driver chose not to implement absolute()?
Are there any workaround for the limitation?
Thanks for the hint from Mark Rotteveel. Now I understand better and let me post an answer to my own question.
Implementation of absolute() is optional
As specified by the Interface of ResultSet#absolute() (link), the implementation for absolute() is optional -- especially when the result set type is TYPE_FORWARD_ONLY.
Workaround
In my case, the result set comes from a Spark Thrift Server (STS) so I guess it is indeed forward-only. So the question became how to instruct my application to NOT making a call to absolute(), which is basically for cursor movement.
SmartGWT-specific answer
For SmartGWT, this is controlled by a property called sqlPaging, which we can specified for an OperationBinding. The right value to use seems to be dropAtServer (more reference here). So I set my SmartGWT DataSource XML file to something like this
<operationBindings>
<operationBinding operationType="fetch" progressiveLoading="false"
sqlPaging="dropAtServer"
>
After that I saw another error, which is now related to HiveConnection#commit():
java.sql.SQLException: Method not supported
at org.apache.hive.jdbc.HiveConnection.commit(HiveConnection.java:742)
at org.apache.commons.dbcp.DelegatingConnection.commit(DelegatingConnection.java:334)
at com.isomorphic.sql.SQLTransaction.commitTransaction(SQLTransaction.java:307)
at com.isomorphic.sql.SQLDataSource.commit(SQLDataSource.java:4673)
After more digging, I realized that the right property for SmartGWT to control the commit behavior is autoJoinTransactions and I should set it to false (more reference here). After these two changes, I could get may application to talk to STS via jdbc.HiveDriver
For anyone out there who are also trying this, here is my full settings for the driver in SmartGWT's server.properties (more reference here)
sql.defaultDatabase: perf2 # this name is picked by me, but it can be anyname
sql.perf2.driver.networkProtocol: tcp
sql.perf2.driver: org.apache.hive.jdbc.HiveDriver # important
sql.perf2.database.type: generic # important
sql.perf2.autoJoinTransactions: false # important
sql.perf2.interface.type: driverManager # important
sql.perf2.driver.url: jdbc:hive2://host:port # important -- pick your host:port
sql.perf2.driver.user: someuser # important -- pick your username
sql.perf2.interface.credentialsInURL: true
sql.perf2.driver.databaseName: someDb
sql.perf2.driver.context:
Related
I have a collection of Person, stored in a legacy mongodb server (2.4) and accessed with the mongoid gem via the ruby mongodb driver.
If I perform a
Person.where(email: 'some.existing.email#server.tld').first
I get a result (let's assume I store the id in a variable called "the_very_same_id_obtained_above")
If I perform a
Person.find(the_very_same_id_obtained_above)
I got a
Mongoid::Errors::DocumentNotFound
exception
If I use the javascript syntax to perform the query, the result is found
Person.where("this._id == #{the_very_same_id_obtained_above}").first # this works!
I'm currently trying to migrate the data to a newever version. Currently mongodbrestore-ing on amazon documentdb to make tests (mongodb 3.6 compatible) and the issue remains.
One thing I noticed is that those object ids are peculiar:
5ce24b1169902e72c9739ff6 this works anyway
59de48f53137ec054b000004 this requires the trick
The small number of zeroes toward the end of the id seems to be highly correlated with the problem (I have no idea of the reason).
That's the default:
# Raise an error when performing a #find and the document is not found.
# (default: true)
raise_not_found_error: true
Source: https://docs.mongodb.com/mongoid/current/tutorials/mongoid-configuration/#anatomy-of-a-mongoid-config
If this doesn't answer your question, it's very likely the find method is overridden somewhere in your code!
Overflowers
Please pardon my question if it's answer it or the answer is naive.
I have a very basic Spring Boot (1.5.4) logging setup in application.properties:
logging.level.org=WARN
logging.level.com=WARN
logging.level.springfox=OFF
logging.level.org.hibernate.hql.internal.ast=ERROR
logging.level.com.MyCompany.kph=DEBUG
logging.file=/var/MyProduct/logs/MyProduct.log
logging.file.max-size=2GB
logging.file.max-history=100
The 2GB is not being honoured. No value I put in there is being honoured. Even xxxxx as a value does not cause a blow-up.
logging.file does - and I can see that being used inside DefaultLogbackConfiguration.
From my source-following I can see method DefaultLogbackConfiguration#setMaxFileSize(a, b) being called. But that method is fixed at 10MB. This aligns with the behaviour i'm seeing.
Am I doing something wrong and triggering the very default behaviour? Or Does default behavior get loaded first then specific stuff goes on top? (If it does, I can't find it and it's not working for me).
Can someone point to me where max-size gets consumed and used?
Thanks
Rich
Christ just by writing this post and reading the docs for MY-SPRING-VERSION, I see max-size is not used at all. That is why it's not working.
https://docs.spring.io/spring-boot/docs/1.5.19.BUILD-SNAPSHOT/reference/htmlsingle/#boot-features-logging
I have written a beanshell sampler which has the objective of setting the multiple values for a single key. I tried with lpush just for the start. Then I went on using the zadd method which I believe is for adding multiple values for a single key.
But, in both the cases, I am getting an error in the log viewer that the respective method in the jedis class does not exist. How do I go about resolving this error.
I have tried to replace the jedis jar file(current version i am using is: 2.2.1) with the latest version(in the lib folder) in order to ensure the completeness of the methods to be used. But then, while starting the jmeter from command line throws java.lang.VerifyError.
Please let me know the solution for this problem.
Show your code so that we can help more precisely.
Another option is to use this if you only want to read the values:
http://jmeter-plugins.org/wiki/RedisDataSet/
Curl has an option which allows me to specify to which IP a domain should be resolved
e.g. curl --resolve example.com:443:1.2.3.4 https://example.com/foo
to make sure that a very specific server is called
(e.g. when multiple servers have the same vhost with a load balancer usually in front of it and there are multiple applications running on the same port with different vhosts)
How do I set this value when using Ethon? https://github.com/typhoeus/ethon
This is how I'd expect it to work
Ethon::Easy.new(url: "https://example.com/foo", :resolve => "example.com:443:1.2.3.4")
but I'm getting an invalid value exception (I have tried multiple different formats that came to mind)
Ethon::Errors::InvalidValue: The value: example.com:443:1.2.3.4 is invalid for option: resolve.
I took a look at the code but couldn't figure out how I'd have to provide the value - and the documentation on this is a bit scarce
Thanks in advance for any reply that might point me in the right direction
Thanks to i0rek on Github I got the answer I was looking for:
resolve = nil
Ethon::Curl.slist_append(resolve, "example.com:443:1.2.3.4")
e = Ethon::Easy.new(url: "https://example.com/foo", :resolve => resolve)
#=> #<Ethon::Easy:0x007faca2574f30 #url="https://example.com/foo", ...>
Further information can be found here:
https://github.com/typhoeus/ethon/issues/95#event-199961240
Problem:
Attempting to use the JYTHON command below and I cannot retrieve the id of my active specification defined at a node-server level in Websphere. I believe its a syntax issue but I'm not sure what.
Code:
AdminConfig.getid('/Cell:mycell/Node:mynode/Server:myserver/J2CActivationSpec:myActiveSpecName/')
Problem Notes:
I do not get a invalid object error so I believe I have the syntax right but it just cannot find the resource even though it exists.
I am using the AdminConfig.getid() as a way to check if the resource already exists in order to do a modify or a create.
If I use the following code: AdminConfig.getid('/J2CActivationSpec:myActiveSpecName/') it will find it but not if I use a more specific path listed above.
Reference Material:
IBM Documentation
Containment paths are always a little tricky. In my (limited) experience, even if you can trace the path by AdminConfig.parents, you may not always be able to use getid.
Are you restricted to using getid? If not, here are some alternatives that will get you an ActivationSpec at the /Cell/Node/Server level:
Querying using AdminConfig.list
This approach will list the Activation Specifications at the specified scope (in our case, the server), and grab the one that has it's name attribute equal to 'myActiveSpecName'.
server = AdminConfig.getid('/Cell:mycell/Node:mynode/Server:myserver')
activationSpec = ''
for as in AdminConfig.list('J2CActivationSpec', server).splitlines():
if AdminConfig.showAttribute(as, 'name') == 'myActiveSpecName'
activationSpec = as
print 'found it :)'
Using Wildcards
This approah uses AdminConfig.list as well, but with a pattern to narrow down your list. If you know your activation spec's configuration begins with myActiveSpecName, then you can do the following:
activationSpec = AdminConfig.list('J2CActivationSpec', 'myActiveSpecName*')