基于prometheus与grafana搭建fsimage解析监控（1）

原创

Bob hadoop

修改于 2021-03-11 16:07:30

8320

修改于 2021-03-11 16:07:30

文章被收录于专栏：日常杂记

1、获取fsimage信息，对于超级小的集群，或者是文件数较少的集群可以用命令获取。

[root@whx nn]# su - hdfs
[hdfs@whx ~]$ hdfs dfsadmin -fetchImage  /data01/dfs/nn/
21/03/10 14:53:32 INFO namenode.TransferFsImage: Opening connection to http://b10.whx.com:50070/imagetransfer?getimage=1&txid=latest
21/03/10 14:53:32 INFO namenode.TransferFsImage: Image Transfer timeout configured to 60000 milliseconds
21/03/10 14:53:32 INFO namenode.TransferFsImage: Transfer took 0.04s at 885.71 KB/s

对于生产中较大的集群我们就不弄用上面的方法获取fsimage，选择去nn目录下的current目录下的fsimage最新的即可，缺点是并不及时。

-rw-r--r-- 1 hdfs hdfs   95271 3月  11 06:58 fsimage_0000000000000123318
-rw-r--r-- 1 hdfs hdfs      62 3月  11 06:58 fsimage_0000000000000123318.md5
-rw-r--r-- 1 hdfs hdfs  101002 3月  11 07:58 fsimage_0000000000000124989
-rw-r--r-- 1 hdfs hdfs      62 3月  11 07:58 fsimage_0000000000000124989.md5
-rw-r--r-- 1 hdfs hdfs       7 3月  11 08:22 seen_txid
-rw-r--r-- 1 hdfs hdfs     172 3月   3 14:39 VERSION
[root@whx current]# pwd
/data01/dfs/nn/current

2、进行格式化fsimage变为可读性文件

[hdfs@whx nn]$ hdfs oiv -i /data01/dfs/nn/fsimage_0000000000000095032 -o /data01/dfs/nn/fsimage.csv -p Delimited  -delimiter ","
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Loading string table
21/03/10 14:56:04 INFO offlineImageViewer.FSImageHandler: Loading 5 strings
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Loading inode references
21/03/10 14:56:04 INFO offlineImageViewer.FSImageHandler: Loading inode references
21/03/10 14:56:04 INFO offlineImageViewer.FSImageHandler: Loaded 0 inode references
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Loading directories
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Loading directories in INode section.
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Found 46 directories in INode section.
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Finished loading directories in 22ms
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Loading INode directory section.
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Scanned 38 INode directories to build namespace.
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Finished loading INode directory section in 3ms
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Found 441 INodes in the INode section
21/03/10 14:56:04 INFO offlineImageViewer.PBImageTextWriter: Outputted 441 INodes.

上面这个步骤对于较小的fsimage可以，但是当fsimage较大时，在hadoop客户端解析的时候会报full gc。需要我们改变HADOOP_HEAPSIZE

可以在执行上面的命令前执行如下

[root@whx current]# export  HADOOP_HEAPSIZE=15240

参数的意义调大hadoop执行内存，单位为M

3、将fsimage.csv文件转化一下。删除fsimage的首行表头。查看列头信息。

[hdfs@whx nn]$ ls
current  fsimage_0000000000000095032  fsimage.csv  in_use.lock
[hdfs@whx nn]$ sed -i -e "1d" /data01/dfs/nn/fsimage.csv
[root@whx nn]# head -n1 fsimage.csv
Path,Replication,ModificationTime,AccessTime,PreferredBlockSize,BlocksCount,FileSize,NSQUOTA,DSQUOTA,Permission,UserName,GroupName

4、进入hive建立fsimage解析表

[hdfs@whx nn]$ hive
Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=512M; support was removed in 8.0
Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=512M; support was removed in 8.0
Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=512M; support was removed in 8.0
Logging initialized using configuration in jar:file:/opt/cloudera/parcels/CDH-5.12.1-1.cdh5.12.1.p0.3/jars/hive-common-1.1.0-cdh5.12.1.jar!/hive-log4j.properties
WARNING: Hive CLI is deprecated and migration to Beeline is recommended.
hive> show databases;
OK
default
Time taken: 1.71 seconds, Fetched: 1 row(s)
hive> create database mydb;
OK
Time taken: 0.132 seconds
hive> use mydb;
OK
Time taken: 0.029 seconds

hive> CREATE TABLE `fsimage_info_csv`(
    >   `path` string,
    >   `replication` int,
    >   `modificationtime` string,
    >   `accesstime` string,
    >   `preferredblocksize` bigint,
    >   `blockscount` int,
    >   `filesize` bigint,
    >   `nsquota` string,
    >   `dsquota` string,
    >   `permission` string,
    >   `username` string,
    >   `groupname` string)
    > ROW FORMAT SERDE
    >   'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'
    > WITH SERDEPROPERTIES (
    >   'field.delim'=',',
    >   'serialization.format'=',')
    > STORED AS INPUTFORMAT
    >   'org.apache.hadoop.mapred.TextInputFormat'
    > OUTPUTFORMAT
    >   'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'
    > LOCATION
    >     'hdfs://whx/user/hive/warehouse/fsimage_info_csv';
OK
Time taken: 0.177 seconds

建表语句：基本注意点表明fsimage_info_csv

CREATE TABLE `fsimage_info_csv`(
  `path` string, 
  `replication` int, 
  `modificationtime` string, 
  `accesstime` string, 
  `preferredblocksize` bigint, 
  `blockscount` int, 
  `filesize` bigint, 
  `nsquota` string, 
  `dsquota` string, 
  `permission` string, 
  `username` string, 
  `groupname` string)
ROW FORMAT SERDE 
  'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe' 
WITH SERDEPROPERTIES ( 
  'field.delim'=',', 
  'serialization.format'=',') 
STORED AS INPUTFORMAT 
  'org.apache.hadoop.mapred.TextInputFormat' 
OUTPUTFORMAT 
  'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'
LOCATION
    'hdfs://whx/user/hive/warehouse/fsimage_info_csv'

上传fsimage表信息，遇到了权限问题，赋权解决。

[hdfs@whx ~]$ hdfs dfs -put /data01/dfs/nn/fsimage.csv /user/hive/warehouse/fsimage_info_csv01/
put: /data01/dfs/nn/fsimage.csv (Permission denied)
[hdfs@whx ~]$ logout
您在 /var/spool/mail/root 中有新邮件
[root@whx nn]# chown -R hdfs:hdfs /data01/
[root@whx nn]# su - hdfs
上一次登录：四 3月 11 08:46:47 CST 2021pts/0 上
[hdfs@whx ~]$ hdfs dfs -put /data01/dfs/nn/fsimage.csv /user/hive/warehouse/fsimage_info_csv01/
[hdfs@whx ~]$

查hdfs的3级目录语句，可以在hive中执行，注意修改表名

原创声明：本文系作者授权腾讯云开发者社区发表，未经许可，不得转载。

如有侵权，请联系 cloudcommunity@tencent.com 删除。

node.js