Downloaded sandbox; installed in VMWare ESX environment, came up without a hitch.
I'm now wading through the "Hadoop Tutorial - Getting Started with HDP" and have run into an issue in Lab 2, Step 2.
I pasted in the query text to define the tables, kicked off execution, and it sat there "running" a long time. You could even see it in a "running" state in the History tab. I let it "run" for several minutes and was concerned that a simple table definition should take so long. So I attempted to stop execution, which had no effect at all. After about 20 minutes or so, I rebooted the sandbox.
Once the sandbox was back up, I went back into Hive and attempted to define the table again. This time the query said it was "running" again and sat there for minutes again. I got sick of waiting this second time and attempted to stop execution and boom... error message pops up on the upper right telling me the script failed because the table already exists. Um... ok...
I look in the database tree and yep, there's the table created by a script that never appeared to finish and should've been killed by a reboot.
I shrugged it off and went to create the second table. Same deal. Run table creation query and it will just sit there... for hours if I let it (I did). I try to kill it, no apparent change. I reboot the sandbox, no apparent change. I try to run the table create again, no apparent change. I try to kill the table create a second time and boom... error saying the script failed because the table's now suddenly there.
What gives? I'd like to hand this off to some other personnel to start testing how we can integrate the Hadoop stack into various things around here but it's not a good sign when the canned sandbox is doing odd things like this.
Hello David, how much RAM do you have in your environment? The 2.4 version of the sandbox needs 8GB.
Also more detailed information can be found in the hive logs, if you're still having issues please attach:
I'm having the same issue.
The hive.log states: SLF4J: Class path contains multiple SLF4J bindings. SLF4J: Found binding in [jar:file:/usr/hdp/18.104.22.168-2950/hadoop/lib/slf4j-log4j12-1.7.10.jar!/org/slf4j/impl/StaticLoggerBinder.cl ass] SLF4J: Found binding in [jar:file:/usr/hdp/22.214.171.124-2950/spark/lib/spark-assembly-126.96.36.199.3.2.0-2950-hadoop188.8.131.52.3.2.0-2950.jar! /org/slf4j/impl/StaticLoggerBinder.class] SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation. SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]
http://www.slf4j.org/codes.html#multiple_bindings states this is only a warning and not an actual error. However when doing a search on the "SLF4J: Class path contains multiple SLF4J bindings.", this has caused major issues.
The main issue I'm having is when I go into the Hive History, I see the query is still running. I cannot find a job running when I use the "mapred job -list" command.
@David Van Tries
I am having the same issue. The queries to create the table never finished. I had to stop execution. For one of them, it seemed to stop. For the other, the "stop" never finished. The history shows the 2nd query still running (for 8 minutes now).
When I try to re-run either of the queries, they fail, telling me that the table already exists. when I refresh the browser and look through the default database, neither table is there. sort of like the twilight zone. even though neither finished and I seemed to be successful in killing the first. I can't create the tables now, and the aren't visible.
running on a 16 GB 8-way machine. so this should be enough power to create a table I hope. I have looked in the log files and nothing there.
I was working with another person encountering this exact issue and we went back and forth looking for a solution. They were finally able to resolve it when they switched from Internet Explorer to Firefox. I don't know for sure if the browser was the root issue, but seeing that this hasn't been resolved, it is worth a try.
Same deal, ran the 1st DDL, it hung, ran the 2nd, it hung but was able to kill the 1st. Tried to kill the 2nd job but no bueno. Tried restarting hive as well as restarting ALL services and that did not get rid of it either. Also tried killing from chrome as well as firefox and that had no effect either.
However, if I change the statement to this:
create table geolocation stored as ORC tblproperties ("orc.compress"="ZLIB") as select * from geolocation_stage ;
it works..... even though the other statement continues to remain in a RUNNING state.