Re: hiveserver2 with OpenLDAP ?

2013-08-23 Thread Mikhail Antonov
You don't need to rebuild Hive, no. I see 2 solutions for that - 1) modify your LDAP data schema so it accepts uid instead of cn attribute (and actually I think it should be working, as after all there can't be so obvious bug in Hive...), so probably whoever tested that provider was using differe

Re: hiveserver2 with OpenLDAP ?

2013-08-23 Thread Sanjay Subramanian
Thanks a lot Mikhail for getting back. That means I cannot use this using beeline unless I change the code and build hive again ? Thanks sanjay From: Mikhail Antonov mailto:olorinb...@gmail.com>> Reply-To: "user@hive.apache.org" mailto:user@hive.apache.org>> Date:

Re: hiveserver2 with OpenLDAP ?

2013-08-23 Thread Mikhail Antonov
To validate what I said, run hiveserver2 in debug mode, put breakpoint in this class, authenticate method, and see if the exception is being thrown. 2013/8/23 Mikhail Antonov > I see the same behavior and here's the reason. > > LdapAuthenticationProviderImpl - that the one responsible for LDAP

Re: hiveserver2 with OpenLDAP ?

2013-08-23 Thread Mikhail Antonov
I see the same behavior and here's the reason. LdapAuthenticationProviderImpl - that the one responsible for LDAP authentication in Hive. Look at this class. It has snippet (CDH 4.2.1, hive 0.10): // setup the security principal String bindDN; if (baseDN != null) { bindDN = "uid="

Re: DISTRIBUTE BY works incorrectly in Hive 0.11 in some cases

2013-08-23 Thread Shahansad KP
I think the problem lies with in the group by operation. For this optimization to work the group bys partitioning should be on the column 1 only. It wont effect the correctness of group by, can make it slow but int this case will fasten the overall query performance. On Fri, Aug 23, 2013 at 5:55

hiveserver2 with OpenLDAP ?

2013-08-23 Thread Sanjay Subramanian
Hi guys I tested hiveserver2 with Active directory - It works With Open LDAP it does not Is there any specific syntax for specifying the LDAP url or baseDN ? hive.server2.authentication.ldap.url ldap://myserver.corp.nextag.com:389 hive.server2.authentication.ldap.baseDN dc=wizetest,dc

Re: DISTRIBUTE BY works incorrectly in Hive 0.11 in some cases

2013-08-23 Thread Pala M Muthaia
I have attached the hive 10 and 11 query plans, for the sample query below, for illustration. On Fri, Aug 23, 2013 at 5:35 PM, Pala M Muthaia wrote: > Hi, > > We are using DISTRIBUTE BY with custom reducer scripts in our query > workload. > > After upgrade to Hive 0.11, queries with GROUP BY/DI

DISTRIBUTE BY works incorrectly in Hive 0.11 in some cases

2013-08-23 Thread Pala M Muthaia
Hi, We are using DISTRIBUTE BY with custom reducer scripts in our query workload. After upgrade to Hive 0.11, queries with GROUP BY/DISTRIBUTE BY/SORT BY and custom reducer scripts produced incorrect results. Particularly, rows with same value on DISTRIBUTE BY column ends up in multiple reducers

Re: Hive 0.11 with Cloudera CHD4.3 MR v1

2013-08-23 Thread Justin Workman
I am not using yarn either, but this value MUST be set. It is one of the 2 issues I ran into. Something to do with the path to the intermediate files not being correct for MR1, this sets that value to a bogus value. I have mine set as yarn.resourcemanager.address foobar In my hive-site.xml.

Re: Hive 0.11 with Cloudera CHD4.3 MR v1

2013-08-23 Thread Jim Colestock
I miss the bit in the middle about adding the value to hive-site.xml (did the skip ahead bit).. So after adding: yarn.resourcemanager.address foobar to hive-site.xml, It's working! Thanks so much for your help.. have a great weekend.. JC On Aug 23, 2013, at 1:37 PM, Justin Workma

Re: Hive 0.11 with Cloudera CHD4.3 MR v1

2013-08-23 Thread Jim Colestock
I'm not using Yarn, But I'll give that a shot.. Is hive 11 dependent on Yarn? JC On Aug 23, 2013, at 1:19 PM, Justin Workman wrote: > This looks similar to the problem you get when you don't set > yarn.resourcemanager.address. > > Sent from my iPhone > > On Aug 23, 2013, at 11:12 AM, Ji

Re: Hive 0.11 with Cloudera CHD4.3 MR v1

2013-08-23 Thread Justin Workman
I just went through this same process again and deployed to another 5 node cluster, and it is also working as expected. I have ran anything very intensive yet, but did spawn several mapreduce jobs. On Fri, Aug 23, 2013 at 11:11 AM, Jim Colestock wrote: > Hello Justin, > > Thanks so much for the

Re: Hive 0.11 with Cloudera CHD4.3 MR v1

2013-08-23 Thread Justin Workman
This looks similar to the problem you get when you don't set yarn.resourcemanager.address. Sent from my iPhone On Aug 23, 2013, at 11:12 AM, Jim Colestock wrote: Hello Justin, Thanks so much for the write up.. I'm compiling now.. I did find one error in your instructions: patch -p0 > HIVE-4

Re: Hive 0.11 with Cloudera CHD4.3 MR v1

2013-08-23 Thread Jim Colestock
Hello Justin, Thanks so much for the write up.. I'm compiling now.. I did find one error in your instructions: patch -p0 > HIVE-4619.D10971.1.path I believe it should be: patch -p0 < HIVE-4619.D10971.1.patch But alas, I still doesn't work for me. I managed to get it to compile after us

Re: Interpreting explain plan in hive

2013-08-23 Thread Edward Capriolo
The only bad plan would be a plan that ends up producing the wrong result, or throwing an exception. generally I look to see if there is a way to have less map/reduce stages, the best plan is usually the one that does the job in a single map reduce, or even better a single map only job. On Fri, A