overview
由于etcd保留了集群中所有版本的确切的历史记录，所以需要定期压缩历史记录以避免性能下降及最终将存储空间耗尽。etcd集群需要定期维护才能保持可靠性，etcd的维护通常可以自动执行，无需停机，也不会影响性能下降。
压缩历史数据会丢弃给定的压缩版本之前所有的历史数据，在压缩数据后，后端数据存储可能会出现内部碎片，碎片仍会占用存储空间，因此需要对碎片进行整理将空间释放回文件系统。
etcd 维护流程
-
备份etcd数据
$ cat etcd-snapshot #etcd-snapshot脚本
#!/bin/bash
BACKUP_PATH=/var/lib/etcd-backup
mkdir -p $BACKUP_PATH
SNAP_FILE="etcd-snapshot-`date +%Y%m%d-%H%M%S`"
. /etc/etcd/etcdrc; /usr/local/bin/etcdctl snapshot save "${BACKUP_PATH}/${SNAP_FILE}"
cd $BACKUP_PATH; tar -czvf ${SNAP_FILE}.tar.gz $SNAP_FILE
cd $BACKUP_PATH; find ./ -mtime +7 -type f -delete
rm -f ${BACKUP_PATH}/${SNAP_FILE}
$ sh etcd-snapshot #执行脚本备份
2019-01-22 18:28:41.665959 I | warning: ignoring ServerName for user-provided CA for backwards compatibility is deprecated
Snapshot saved at /var/lib/etcd-backup/etcd-snapshot-20190122-182841
etcd-snapshot-20190122-182841
-
获取当前etcd集群所有member endpoints
- 导入ENV
$ cat etcdrc export ETCDCTL_API=3 export ETCDCTL_ENDPOINTS="127.0.0.1:2379" export ETCDCTL_CACERT=/etc/etcd/ssl/ca.crt export ETCDCTL_CERT=/etc/etcd/ssl/etcd.crt export ETCDCTL_KEY=/etc/etcd/ssl/etcd.key $ source etcdrc # 执行- 获取member endpoints
$ etcdctl member list b647a4c5532e517, started, vms70322, https://10.5.xx.93:2380, https://10.5.xx.93:2379 c7eeae8975872e8, started, vms70324, https://10.5.xx.95:2380, https://10.5.xx.95:2379 29a57fa29bbcf1a7, started, vms70323, https://10.5.xx.94:2380, https://10.5.xx.94:2379 caedcb6700c3f637, started, vms70321, https://10.5.xx.92:2380, https://10.5.xx.92:2379 f5c43730caf9ca87, started, vms70320, https://10.5.xx.91:2380, https://10.5.xx.91:2379 -
获取当前最新的etcd revision版本$rev_num
$ etcdctl endpoint status --write-out="json" | egrep -o '"revision":[0-9]*' | egrep -o '[0-9].*'
39042173
-
保留最新的100个revision版本，压缩($rev_num -100)旧版本数据
$ etcdctl --command-timeout=300s compact 39042073
-
碎片整理（对每个member逐一进行，先非leader，再leader）
$ etcdctl --command-timeout=300s defrag --endpoints=https://10.5.xx.xx:2379
Finished defragmenting etcd member[https://10.xx.xx.xxx:2379]
-
维护结果
数据压缩完成之后数据大小DB size将降低：
DBSize.png
disk.png
