sorry， I forgot it.
It's because of the jvm method Stack .It was 128k before. we were sending
hiveserver concurrent and max jobnum would be 10. Most of These jobs used udf
to work, and udf used more mem because of recurrencion. Now we set Xss 512k,
and hiveserver has been nomal.
在 2011-12-22
这个原因知道了 是因为hiveserver jvm 方法栈设置低，之前设置128k，但是hiveserver里udf 类的函数需要递归
等，需求较大，修正为512k，已经稳定运行3天。
At 2011-12-14 10:21:53,"王锋" wrote:
when hiveserver is executing sql, it throws an exception:
2011-12-14 00:40:43,762 Stage-1 map = 90%, reduce = 13%
2011-12-14 00:40:45,771 Stage-1 map = 92%, reduce
Hi,
It looks like at one point there was an interest in supporting common
table expressions in Hive.
https://issues.apache.org/jira/browse/HIVE-1180
Is this still a planned feature?
ye, I will be looking at this question.
在 2011-12-21 21:49:31，"北斗七" 写道：
>i aslo meet the same question,so if you know how to slove it ,please
>tell me .Ths.
>
>在 2011年12月21日 下午6:00，王锋 写道：
>> I just create a table fool; And table data is :
>> sellerId consumerId
>> 1000,4000
>> 1000,4001
>>
Hi all,
I just found that before calling a "MAP ... FROM ... USING resource" command ,
the resource need first to be added to every machine in the cluster. Actually
this makes perfect sense to me but I thought naively that Hive could do this
automatically for me...
Anyway, I found in the docum
i aslo meet the same question,so if you know how to slove it ,please
tell me .Ths.
在 2011年12月21日 下午6:00，王锋 写道：
> I just create a table fool; And table data is :
> sellerId consumerId
> 1000,4000
> 1000,4001
> 1002,4002
> 1003,4000
> 1004,4001
> 1005,4002
> 1000,4000
>
> total 8 rows;
>
> then I e
Thanks for yours .
在 2011-12-21 17:24:24，"sangeetha s" 写道：
:)
On Tue, Dec 20, 2011 at 2:56 AM, John Sichi wrote:
Carl, big thanks to you as release manager for pushing this one through!
JVS
On Dec 19, 2011, at 10:29 AM, Carl Steinbach wrote:
> The Apache Hive team is proud to announce t
I just create a table fool; And table data is :
sellerId consumerId
1000,4000
1000,4001
1002,4002
1003,4000
1004,4001
1005,4002
1000,4000
total 8 rows;
then I execute sql "select consumerId from fool".
When the job was executing, I watch one map log. through "http://ip:50030/";
syslog
:)
On Tue, Dec 20, 2011 at 2:56 AM, John Sichi wrote:
> Carl, big thanks to you as release manager for pushing this one through!
>
> JVS
>
> On Dec 19, 2011, at 10:29 AM, Carl Steinbach wrote:
>
> > The Apache Hive team is proud to announce the the release of Apache
> > Hive version 0.8.0.
> >
>
