You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@kylin.apache.org by cheney <53...@qq.com> on 2018/12/06 12:52:22 UTC
回复: Error size statis of segment
kylin version is 2.4.1 and kylin metadata is stored in hbase. hbase version is hbase-1.2.0-cdh5.10.1
------------------ 原始邮件 ------------------
发件人: "邓林钢"<ch...@gmail.com>;
发送时间: 2018年12月6日(星期四) 晚上8:09
收件人: "user"<us...@kylin.apache.org>;
主题: Re: Error size statis of segment
Hi,cheney, What`s your kylin version and kylin metadata type?
cheney <53...@qq.com> 于2018年12月6日周四 下午5:49写道:
Hi, guys
Sometimes, the size of new segment is always less than 1MB after building job. In fact, size of new htable in hdfs is larger than 1MB. I find the result of "regionLoad.getStorefileSizeMB()" always is 0. I think it is unreasonable calculating segment size by "regionLoad.getStorefileSizeMB()" method. Is it a bug?
1. kylin monitor show size of segment is as followings:
2. hdfs show size of HBase Table: KYLIN_RF4S1QAHE4
回复: Error size statis of segment
Posted by cheney <53...@qq.com>.
I haved created issue KYLIN-3713 in jira.
------------------ 原始邮件 ------------------
发件人: "Yichen Zhou"<zh...@gmail.com>;
发送时间: 2018年12月7日(星期五) 下午2:58
收件人: "user"<us...@kylin.apache.org>;
主题: Re: Error size statis of segment
Hi Cheney,
I also observed this issue.
I think it is a bug in Kylin. Could you please create a ticket on JIRA?
Thanks & Regards,
Yichen
On Thu, Dec 6, 2018 at 8:52 PM cheney <53...@qq.com> wrote:
kylin version is 2.4.1 and kylin metadata is stored in hbase. hbase version is hbase-1.2.0-cdh5.10.1
------------------ 原始邮件 ------------------
发件人: "邓林钢"<ch...@gmail.com>;
发送时间: 2018年12月6日(星期四) 晚上8:09
收件人: "user"<us...@kylin.apache.org>;
主题: Re: Error size statis of segment
Hi,cheney, What`s your kylin version and kylin metadata type?
cheney <53...@qq.com> 于2018年12月6日周四 下午5:49写道:
Hi, guys
Sometimes, the size of new segment is always less than 1MB after building job. In fact, size of new htable in hdfs is larger than 1MB. I find the result of "regionLoad.getStorefileSizeMB()" always is 0. I think it is unreasonable calculating segment size by "regionLoad.getStorefileSizeMB()" method. Is it a bug?
1. kylin monitor show size of segment is as followings:
2. hdfs show size of HBase Table: KYLIN_RF4S1QAHE4
Re: Error size statis of segment
Posted by Yichen Zhou <zh...@gmail.com>.
Hi Cheney,
I also observed this issue.
I think it is a bug in Kylin. Could you please create a ticket on JIRA?
Thanks & Regards,
Yichen
On Thu, Dec 6, 2018 at 8:52 PM cheney <53...@qq.com> wrote:
>
> kylin version is 2.4.1 and kylin metadata is stored in hbase. hbase
> version is hbase-1.2.0-cdh5.10.1
>
> ------------------ 原始邮件 ------------------
> *发件人:* "邓林钢"<ch...@gmail.com>;
> *发送时间:* 2018年12月6日(星期四) 晚上8:09
> *收件人:* "user"<us...@kylin.apache.org>;
> *主题:* Re: Error size statis of segment
>
> Hi,cheney,
> What`s your kylin version and kylin metadata type?
>
> cheney <53...@qq.com> 于2018年12月6日周四 下午5:49写道:
>
>> Hi, guys
>>
>> Sometimes, the size of new segment is always less than 1MB after
>> building job. In fact, size of new htable in hdfs is larger than 1MB. I
>> find the result of "regionLoad.getStorefileSizeMB()" always is 0. I
>> think it is unreasonable calculating segment size by "
>> regionLoad.getStorefileSizeMB()" method. Is it a bug?
>>
>> * 1. kylin monitor show size of segment is as followings:*
>>
>>
>> * 2. hdfs show size of HBase Table: KYLIN_RF4S1QAHE4*
>>
>>
>>
>>