一、cassandra的备份和恢复
cassandra的备份恢复主要是通过snapshots 来实现。
步骤：
备份阶段：
1. 生成快照；
恢复阶段：
1. 清空表的数据（truncate table tablename）或者创建表结构；
2. 把各个表复制到对应标的数据目录下，覆盖原来的数据；
3. refresh 加载恢复数据；
1、生成快照(snapshots)
语法：
nodetool -h 服务器ip -p 端口号 snapshots 数据库名 #全库快照
nodetool -h 服务器ip -p 端口号 snapshots -t 快照名称 -kt 数据库名.表名 #某个表快照
注：不指定-t 会自动生成时间戳字符串
生成好的snapshots 保存在数据目录对应的表目录下的snapshots，例:table-uuid/snapshots/snapshotsname
[email protected]:~# nodetool -h localhost -p 7199 snapshot -t xn_dolphin_1-20181010 xn_dolphin_1
Requested creating snapshot(s) for [xn_dolphin_1] with snapshot name [xn_dolphin_1-20181010] and options {skipFlush=false}
Snapshot directory: xn_dolphin_1-20181010
[email protected]:~# nodetool -h localhost -p 7199 snapshot xn_dolphin_1
Requested creating snapshot(s) for [xn_dolphin_1] with snapshot name [1539180816386] and options {skipFlush=false}
Snapshot directory: 1539180816386
[email protected]:~# date
Wed Oct 10 14:14:45 UTC 2018
[email protected]:~# nodetool listsnapshots
Snapshot Details:
Snapshot name Keyspace name Column family name True size Size on disk
1539180816386 xn_dolphin_1 dolphin_conversation_result 5.1 MiB 5.1 MiB
1539180816386 xn_dolphin_1 dolphin_conversation_member 0 bytes 1.02 KiB
1539180816386 xn_dolphin_1 dolphin_wchat_openid 0 bytes 895 bytes
1539180816386 xn_dolphin_1 zoogate_login_info 0 bytes 1.02 KiB
1539180816386 xn_dolphin_1 dolphin_conversation_list 0 bytes 946 bytes
1539180816386 xn_dolphin_1 dolphin_leaving_msg 0 bytes 1.27 KiB
1539180816386 xn_dolphin_1 dolphin_conversation 0 bytes 1.1 KiB
1539180816386 xn_dolphin_1 dolphin_member_inout 0 bytes 1.05 KiB
1539180816386 xn_dolphin_1 dolphin_conversation_message 0 bytes 1.18 KiB
1539180816386 xn_dolphin_1 zoogate_blacklist 0 bytes 1.01 KiB
1539180816386 xn_dolphin_1 dolphin_conversation_visitorinfo 0 bytes 1.2 KiB
1539180816386 xn_dolphin_1 dolphin_conversation_statistics 0 bytes 1 KiB
xn_dolphin_1-20181010 xn_dolphin_1 dolphin_conversation_result 5.1 MiB 5.1 MiB
xn_dolphin_1-20181010 xn_dolphin_1 dolphin_conversation_member 0 bytes 1.02 KiB
xn_dolphin_1-20181010 xn_dolphin_1 dolphin_wchat_openid 0 bytes 895 bytes
xn_dolphin_1-20181010 xn_dolphin_1 zoogate_login_info 0 bytes 1.02 KiB
xn_dolphin_1-20181010 xn_dolphin_1 dolphin_conversation_list 0 bytes 946 bytes
xn_dolphin_1-20181010 xn_dolphin_1 dolphin_leaving_msg 0 bytes 1.27 KiB
xn_dolphin_1-20181010 xn_dolphin_1 dolphin_conversation 0 bytes 1.1 KiB
xn_dolphin_1-20181010 xn_dolphin_1 dolphin_member_inout 0 bytes 1.05 KiB
xn_dolphin_1-20181010 xn_dolphin_1 dolphin_conversation_message 0 bytes 1.18 KiB
xn_dolphin_1-20181010 xn_dolphin_1 zoogate_blacklist 0 bytes 1.01 KiB
xn_dolphin_1-20181010 xn_dolphin_1 dolphin_conversation_visitorinfo 0 bytes 1.2 KiB
xn_dolphin_1-20181010 xn_dolphin_1 dolphin_conversation_statistics 0 bytes 1 KiB
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91/snapshots/xn_dolphin_1-20181010# pwd
/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91/snapshots/xn_dolphin_1-20181010
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91# ls
backups mc-1-big-Data.db mc-1-big-Filter.db mc-1-big-Statistics.db mc-1-big-TOC.txt
mc-1-big-CompressionInfo.db mc-1-big-Digest.crc32 mc-1-big-Index.db mc-1-big-Summary.db snapshots
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91/snapshots# ls
1539180816386 testdb_bak xn_dolphin_1-20181010
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91/snapshots# cd xn_dolphin_1-20181010/
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91/snapshots/xn_dolphin_1-20181010# ls
manifest.json mc-1-big-Data.db mc-1-big-Filter.db mc-1-big-Statistics.db mc-1-big-TOC.txt
mc-1-big-CompressionInfo.db mc-1-big-Digest.crc32 mc-1-big-Index.db mc-1-big-Summary.db schema.cql
2、删除数据：
cqlsh:xn_dolphin_1> select count(*) from dolphin_conversation_result;
count
-------
53426
(1 rows)
Warnings :
Aggregation query used without partition key
cqlsh:xn_dolphin_1> truncate table dolphin_conversation_result;
#cassandra 在truncate table的时候会自动创建一个截断表的快照，表目录下的文件除了backups snapshots两个目录，其他都会被删除，删除*.db 文件
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91# nodetool listsnapshots
Snapshot Details:
Snapshot name Keyspace name Column family name True size Size on disk
truncated-1539182023411-dolphin_conversation_result xn_dolphin_1 dolphin_conversation_result 5.1 MiB 5.1 MiB
3、#复制快照文件
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91# ls
backups snapshots
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91# cp snapshots/truncated-1539182023411-dolphin_conversation_result/* .
#权限
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91# chown -R cassandra.cassandra *
4、#恢复数据
语法:nodetool -h 服务器 -p 端口 refresh – 数据库名 数据表名
注：port 为7199
#快照恢复是在schma存在的情况下恢复数据，因此确保schema存在
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91# cqlsh -k xn_dolphin_1
Connected to Test Cluster at 127.0.0.1:9042.
[cqlsh 5.0.1 | Cassandra 3.11.3 | CQL spec 3.4.4 | Native protocol v4]
Use HELP for help.
cqlsh:xn_dolphin_1> desc dolphin_conversation_result ;
CREATE TABLE xn_dolphin_1.dolphin_conversation_result (
siteid text,
converid text,
type int,
content text,
customerid text,
deal_content text,
deal_time bigint,
isdeal int,
submit_time bigint,
supplierid text,
PRIMARY KEY (siteid, converid, type)
) WITH CLUSTERING ORDER BY (converid ASC, type ASC)
AND bloom_filter_fp_chance = 0.01
AND caching = {'keys': 'ALL', 'rows_per_partition': 'NONE'}
AND comment = ''
AND compaction = {'class': 'org.apache.cassandra.db.compaction.SizeTieredCompactionStrategy', 'max_threshold': '32', 'min_threshold': '4'}
AND compression = {'chunk_length_in_kb': '64', 'class': 'org.apache.cassandra.io.compress.LZ4Compressor'}
AND crc_check_chance = 1.0
AND dclocal_read_repair_chance = 0.1
AND default_time_to_live = 0
AND gc_grace_seconds = 864000
AND max_index_interval = 2048
AND memtable_flush_period_in_ms = 0
AND min_index_interval = 128
AND read_repair_chance = 0.0
AND speculative_retry = '99PERCENTILE';
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91# ls
backups mc-1-big-CompressionInfo.db mc-1-big-Digest.crc32 mc-1-big-Index.db mc-1-big-Summary.db schema.cql
manifest.json mc-1-big-Data.db mc-1-big-Filter.db mc-1-big-Statistics.db mc-1-big-TOC.txt snapshots
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91# nodetool refresh -- xn_dolphin_1 dolphin_conversation_result
[email protected]:/var/lib/cassandra/data/xn_dolphin_1/dolphin_conversation_result-d9e929d0cc9511e8a7ad6d2c86545d91# cqlsh -k xn_dolphin_1
Connected to Test Cluster at 127.0.0.1:9042.
[cqlsh 5.0.1 | Cassandra 3.11.3 | CQL spec 3.4.4 | Native protocol v4]
Use HELP for help.
cqlsh:xn_dolphin_1> select count(*) from dolphin_conversation_result;
count
-------
53426
(1 rows)
Warnings :
Aggregation query used without partition key
二、sstableloader迁移工具
在cassandra的bin目录下提供了一个sstableloader工具，这个工具专门用于把一个表的sstable文件导入到一个新的集群中。
注意：如果旧的三台迁移新的三台，因为sstableloader迁移的数据只是执行所在节点上的数据，因此需要old1->new1、old2->new2、old3->new3。
1.旧集群环境：表是mykeyspace.mytable。你的数据存一个3个节点组成的集群中，每个节点的数据都存在/opt/data目录下。
2.新集群环境：地址是192.168.31.185， 先在新集群建离相同名字的keyspace和表结构。
3.在旧集群环境里执行：
bin/sstableloader -d 192.168.31.185 -u cassandra -pw cassandra -t 100 /opt/data/mykeyspace/mytable
其中-u是 用户名 -pw是密码 -t是限制流量100M/bps
等所有节点执行完毕，表数据就成功导入到了新的集群中，当然只要机器io和网络条件允许，你可以多个节点并发执行。
示例：
[email protected]:/$ sstableloader -d 192.168.31.185 -u cassandra -pw cassandra -t 100 /opt/bitnami/cassandra/data/data/xn_dolphin_1/dolphin_conversation_result-2c8866e0ce3711e89b4687b65adcf047/
WARN 16:25:38,358 Small commitlog volume detected at /opt/bitnami/cassandra/bin/../data/commitlog; setting commitlog_total_space_in_mb to 4348. You can override this in cassandra.yaml
WARN 16:25:38,377 Small cdc volume detected at /opt/bitnami/cassandra/bin/../data/cdc_raw; setting cdc_total_space_in_mb to 2174. You can override this in cassandra.yaml
WARN 16:25:38,669 Only 6.847GiB free across all data volumes. Consider adding more capacity to your cluster or removing obsolete snapshots
Established connection to initial hosts
Opening sstables and calculating sections to stream
Streaming relevant part of /bitnami/cassandra/data/data/xn_dolphin_1/dolphin_conversation_result-2c8866e0ce3711e89b4687b65adcf047/mc-1-big-Data.db to [/192.168.31.185]
progress: [/192.168.31.185]0:1/1 100% total: 100% 792.200KiB/s (avg: 792.200KiB/s)
progress: [/192.168.31.185]0:1/1 100% total: 100% 0.000KiB/s (avg: 634.013KiB/s)
Summary statistics:
Connections per host : 1
Total files transferred : 1
Total bytes transferred : 5.053MiB
Total duration : 8175 ms
Average transfer rate : 632.865KiB/s
Peak transfer rate : 792.200KiB/s
迁移完成！
脚本：
python备份脚本，适用于k8s、docker化的cassandra数据备份
#-*- coding:utf-8 -*-
# Author:json
import os,re
import subprocess
import datetime
import zipfile
def exeCommd(cmd):
subp = subprocess.getstatusoutput(cmd)
return subp[1]
def copySnapshots(containerId,snapshot_path,backcup_path):
copyCommd = "kubectl cp containerId:snapshot_path {backcup_path}".format(backcup_path=backcup_path)
exeCommd(copyCommd)
def mkdirFile(path):
if os.path.isdir(path):
print('目录已存在！')
else:
os.makedirs(path)
def exeCommd2(cmd):
subp = subprocess.Popen(cmd,shell=True,stdout=subprocess.PIPE,encoding='utf8')
data = subp.stdout.readlines()
return data
def writeFile(path,filename,content):
with open('{path}/{filename}'.format(path=path,filename=filename),'w+',encoding='utf8') as f:
f.write(content)
def zipFile(zipfname,LOCAL_BACKUP_PATH):
try:
z = zipfile.ZipFile(zipfname,'w',zipfile.ZIP_DEFLATED,allowZip64=True)
print ('开始压缩文件')
for dirpath,dirnames,filenames in os.walk(LOCAL_BACKUP_PATH):
for filename in filenames:
z.write(os.path.join(dirpath,filename))
z.close()
print ('备份结束')
except Exception as e:
print (e)
if __name__ == "__main__":
#需要修改的参数：BACKUP_DIR、CONFIG_YAML、CONTAINER_Id、KEYSPACES_NAME
BACKUP_DIR = '/opt/backup/cassandra'
CONFIG_YAML = '/etc/cassandra/cassandra.yaml'
# CONTAINER_Id = "docker ps|grep 'cassandra'|awk -F' ' '{print $1}'"
CONTAINER_Id = "cassandra-0"
# CURRENT_IP = "ip addr | grep 'eth0' |grep 'inet'|awk '{print $2}'|cut -f1 -d'/'"
KEYSPACES_NAME = ['dolphin','im']
CurrmentDate = datetime.datetime.now().strftime('%Y%m%d%H%I%S')
# containerId = exeCommd(CONTAINER_Id)
#创建 snapshot
# CASSANDRA_SNAPSHOT = "docker exec -it {containerId} nodetool snapshot -t {CURRENT_DATE} ".format(CURRENT_DATE=CURRENT_DATE,containerId=containerId)
CASSANDRA_SNAPSHOT = "kubectl exec -it {containerId} nodetool snapshot".format(CURRENT_DATE=CurrmentDate,containerId=CONTAINER_Id)
#查看cassandra数据目录
CASSANDRA_DataFile="kubectl exec -it {CONTAINER_Id} cat {CONFIG_YAML}|grep 'data_file_directories' -A1|head -2|tail -n1".format(CONFIG_YAML=CONFIG_YAML,CONTAINER_Id=CONTAINER_Id)
a,dataPath = exeCommd(CASSANDRA_DataFile).split()
data_list = []
print(CASSANDRA_SNAPSHOT)
#获取快照的名称
Snapshot_Str = exeCommd(CASSANDRA_SNAPSHOT)
re = re.search('\[\d+\]', Snapshot_Str).group()
SnapshotName = re.replace('[', '').replace(']', '')
print(SnapshotName)
try:
for keyspaceName in KEYSPACES_NAME:
# path = "{dataPath}/{keyspacename}".format(dataPath=dataPath,keyspacename=keyspaceName)
keyspacePath = os.path.join(dataPath,'{keyspacename}'.format(keyspacename=keyspaceName))
print('-------------------------',keyspaceName)
CASSANDRA_TABLE = "kubectl exec -it {containerId} ls {path} ".format(path=keyspacePath,
containerId=CONTAINER_Id)
table_uuid = exeCommd(CASSANDRA_TABLE).split('\n')
for tableFileName in table_uuid:
tableName,uuid = tableFileName.split('-')
# snapshot 保存的路径
snapshotPath = "{path}/{tableFileName}/snapshots/{SnapshotName}".format(tableFileName=tableFileName,
SnapshotName=SnapshotName,
path=keyspacePath)
# 备份到本地的路径
bakcupPath = '{backdir}/{keyspacename}/{tableFileName}'.format(backdir=BACKUP_DIR,
tableFileName=tableFileName,
keyspacename=keyspaceName)
print(bakcupPath)
#创建备份目录
mkdirFile('{back_path}'.format(back_path=bakcupPath))
#copy snapshot backup file
CopySnapshot = 'kubectl cp {containerId}:{table_path} {bakcup_path}/{SnapshotName} '.format(table_path=snapshotPath,containerId=CONTAINER_Id,bakcup_path=bakcupPath,SnapshotName=SnapshotName)
print(CopySnapshot,'>>>:',tableFileName)
if exeCommd(CopySnapshot):
writeFile(BACKUP_DIR, '{CurrmentDate}.log'.format(CurrmentDate=CurrmentDate), 'ok')
print('复制成功！')
else:
writeFile(BACKUP_DIR, '{CurrmentDate}.log'.format(CurrmentDate=CurrmentDate), 'on')
except Exception as e:
print(e)
writeFile(BACKUP_DIR, '{CurrmentDate}.log'.format(CurrmentDate=CurrmentDate), e)
zipfname = BACKUP_DIR + '/' + CurrmentDate + CONTAINER_Id + '.zip'
LOCAL_BACKUP_PATH = os.path.join(BACKUP_DIR,'dolphin')
zipFile(zipfname, LOCAL_BACKUP_PATH)
交流群：725450393
![·EverybodyCOOECOOEnumberdayhmmmaneuvertosettleoncondotorayou[图]](data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACwAAAAAAQABAAACAkQBADs=)