昨天写了一篇针对国产数据库免费的文章，然后一些群友就有不同的意见了。这也难怪，毕竟青菜萝卜，各有所爱！
群友的观点
我提到我们MogDB采用了类似Oracle一样的paper License策略，当然我们目的也就是为了做生态。老实说，在目前国内的严峻环境中，采样这个策略，真的需要非常大的勇气。毕竟现在大家都在勒紧裤腰带过日子呢。
首先我要说明一点，Oracle确实从来没有强调免费，但是确实用户可以免费下载，而且下载的版本跟企业版本也并无两样；这就造成了在国内事实上的”免费“。但是这无疑是让Oracle在国内份额如此之高的一个重要原因。
用的人多了，自然而然生态也很好了，开发者、DBA都趋之若鹜了。
当然MogDB数据库也采用了类似策略，我们本意也是想让更多的开发者、更多的DBA人员参与进来，多多学习和测试，给我们反馈宝库意见。
实际上我曾经测过很多国产数据库，想找人要一份安装软件介质都老费劲了，实在是太麻烦。数据库软件本来就是让人用的，藏着掖着干什么呢？
用开源的PostgreSQL不香吗
既然网友提出了这个问题，那么我就来尝试回答一下。
首先我想说PostgreSQL确实这些发展很快，尤其是国内，这几年因为xc的原因，很多尝试都选择了基于PostgreSQL来进行二次开发增强。
当然也确实很多用户都选择了开源的PostgreSQL，尤其是已经运维能力比较强的客户。如果运维能力跟得上，那么用开源软件是完全ok的，我没有任何意见。
但从另外一个角度来讲，现在国内用户选择数据库替换的原因有很多，成本并不是唯一因素。
就拿我们已经完成国产化上线的这些客户来讲，用户需要的数据库产品，不仅仅是能满足他们的要求，而且需要有较高的Oracle、MySQL兼容性，同时后台还有强大的技术支撑能力。
比如某事业单位使用了PostgreSQL来做替换核心Oracle数据库，那么其改造成本还是非常高的，尽管我知道PostgreSQL也有相关的Oracle兼容性插件。实际上我们做的兼容性很多了，这里简单提一提：
从以上5个维度来讲，我们都做了大量兼容，另外还要补充一点，我们也做了大量针对PostgreSQL的兼容。大家所熟知的鼎捷ERP 之前就是基于PostgreSQL开发，目前部分用户已经迁移到了MogDB，运行平稳。
最后我不得不再次吐槽原生PostgreSQL哪些令人糟心的问题,比如事务ID回卷、增量检查点等至今都没有解决。
“免费”等于不好？便宜无好货？
首先我再次重申一下，MogDB从来没有说过免费。但是如果你要白嫖，我们也就当不知道。哈哈。
继续回到这个话题上来，MySQL、PostgreSQL开源这么多年了，你能说他不好么？ 话说openGauss开源至今也有已经4年了，参与者众多，生态繁荣。
我认为对于一个数据库软件，还不好用，得让用户来评价，这里我给大家附上一段今天刚升级的一套测试环境。
[root@mogdb1 ~]
# ptk cluster status -n mog506
[ Cluster State ]
cluster_name : mog506
cluster_state : Normal
database_version : MogDB-5.0.6 (build 8b0a6ca8)
mogha_installed : yes
[ Datanode State ]
cluster_name | id | ip | port | user | nodename | db_role | state | uptime | upstream
---------------+------+---------------+-------+------+----------+---------+--------+-----------------+-----------
mog506 | 6001 | 172.20.22.125 | 25000 | omm2 | dn_6001 | primary | Normal | 4 days 12:42:34 | -
| 6002 | 172.20.22.126 | 25000 | omm2 | dn_6002 | standby | Normal | 4 days 12:42:30 | -
| 6003 | 172.20.22.128 | 25000 | omm2 | dn_6003 | standby | Normal | 4 days 12:42:27 | -
[root@mogdb1 ~]
#
[root@mogdb1 ~]
# ptk cluster -n mog506 upgrade -y -p /root/MogDB-5.0.7-CentOS-x86_64-all.tar.gz
INFO[2024-06-04T17:40:24.535] PTK Version: 1.4.7 release
INFO[2024-06-04T17:40:24.540] new step recorder, current step:
"start"
Tip:
The upgrade process relies on SSH.
Please ensure that the network is always worked
during the upgrade process.
INFO[2024-06-04T17:40:24.540] parsing package from /root/MogDB-5.0.7-CentOS-x86_64-all.tar.gz
INFO[2024-06-04T17:40:32.355] kernel package name: MogDB-5.0.7-CentOS-64bit.tar.gz
INFO[2024-06-04T17:40:32.356] package version: MogDB-5.0.7,92.924,c4707384
INFO[2024-06-04T17:40:32.356] check db version
INFO[2024-06-04T17:40:32.356] big upgrade:
true
INFO[2024-06-04T17:40:32.356] current version: 5.0.6, target version: 5.0.7
INFO[2024-06-04T17:40:32.356] current number: 92.920, target number: 92.924
INFO[2024-06-04T17:40:32.520] version is ok
INFO[2024-06-04T17:40:32.521] prepare all upgrade sql
INFO[2024-06-04T17:40:33.304] extract upgrade_sql.tar.gz successfully
WARN[2024-06-04T17:40:33.304] no sql file found
for upgrade maindb
WARN[2024-06-04T17:40:33.304] no sql file found
for upgrade otherdb
WARN[2024-06-04T17:40:33.304] no sql file found
for rollback maindb
WARN[2024-06-04T17:40:33.304] no sql file found
for rollback otherdb
INFO[2024-06-04T17:40:33.542] create temp dir
INFO[2024-06-04T17:40:33.560] upload upgrade_extensions.tar.gz
INFO[2024-06-04T17:40:33.579] extract upgrade_extensions.tar.gz to dir /opt/mogdb506/tmp/upgrade
INFO[2024-06-04T17:40:33.749] check cluster status
INFO[2024-06-04T17:40:33.952] cluster state is ok
INFO[2024-06-04T17:40:34.069] check db config: enable_stream_replication
INFO[2024-06-04T17:40:34.120] config is ok: enable_stream_replication=on
INFO[2024-06-04T17:40:34.120] check db config: enable_global_syscache
INFO[2024-06-04T17:40:34.168] config is ok: enable_global_syscache=off
INFO[2024-06-04T17:40:34.168] create remote temporary upgrade directory
INFO[2024-06-04T17:40:34.188]
set cluster to
read-only
INFO[2024-06-04T17:40:34.225] begin: wait_xlog_sync
INFO[2024-06-04T17:40:34.225] check disk size
INFO[2024-06-04T17:40:34.384] check and
wait xlog sync
INFO[2024-06-04T17:40:34.437] finished: wait_xlog_sync
INFO[2024-06-04T17:40:34.437] begin: backup_old_files
INFO[2024-06-04T17:40:34.437] backup old app and tool
INFO[2024-06-04T17:41:00.004] finished: backup_old_files
INFO[2024-06-04T17:41:00.004] begin: exec_pre_sql
INFO[2024-06-04T17:41:00.004]
set cluster to
read-write
INFO[2024-06-04T17:41:00.105] call hook script: before_pre_upgrade_script
INFO[2024-06-04T17:41:01.093] checking cluster status
INFO[2024-06-04T17:41:01.626] cluster status is normal
INFO[2024-06-04T17:41:01.626]
set upgrade_mode to 2
INFO[2024-06-04T17:41:01.660] check and turn on support_extended_features
INFO[2024-06-04T17:41:01.747] execute rollback sql
INFO[2024-06-04T17:41:01.747] execute rollback sql on maindb
INFO[2024-06-04T17:41:01.908] execute rollback sql on otherdb: [template1 dbm template0 tpch sysbench user01_db
test], parallel num: 10
INFO[2024-06-04T17:41:02.275] execute upgrade sql
INFO[2024-06-04T17:41:02.275] execute upgrade sql on maindb
INFO[2024-06-04T17:41:02.425] execute upgrade sql on otherdb: [template1 dbm template0 tpch sysbench user01_db
test], parallel num: 10
INFO[2024-06-04T17:41:02.778] execute checkpoint
INFO[2024-06-04T17:41:03.126] call hook script: after_pre_upgrade_script
INFO[2024-06-04T17:41:03.144] finished: exec_pre_sql
INFO[2024-06-04T17:41:03.144] begin: stop_datanode
......
[pg_trgm, pg_prewarm, postgis, whale, wal2json, oracle_fdw, pg_bulkload, pg_repack, mysql_fdw]
INFO[2024-06-04T17:41:24.944] [172.20.22.125][omm2] upload All-Plugins.tar.gz
INFO[2024-06-04T17:41:24.944] [172.20.22.126][omm2] upload All-Plugins.tar.gz
INFO[2024-06-04T17:41:24.944] [172.20.22.128][omm2] upload All-Plugins.tar.gz
INFO[2024-06-04T17:41:25.268] [172.20.22.125][omm2] make user omm2
's dir(s): /opt/mogdb506/tool/plugins
INFO[2024-06-04T17:41:25.382] [172.20.22.125][omm2] extract All-Plugins.tar.gz to dir /opt/mogdb506/tool/plugins
INFO[2024-06-04T17:41:25.387] [172.20.22.128][omm2] make user omm2's dir(s): /opt/mogdb506/tool/plugins
⣽ executing...INFO[2024-06-04T17:41:25.798] [172.20.22.126][omm2] make user omm2
's dir(s): /opt/mogdb506/tool/plugins
INFO[2024-06-04T17:41:25.929] [172.20.22.126][omm2] extract All-Plugins.tar.gz to dir /opt/mogdb506/tool/plugins
INFO[2024-06-04T17:41:26.417] [172.20.22.125][omm2] change /opt/mogdb506/tool/plugins owner to omm2
INFO[2024-06-04T17:41:26.598] [172.20.22.128][omm2] extract All-Plugins.tar.gz to dir /opt/mogdb506/tool/plugins
INFO[2024-06-04T17:41:27.084] [172.20.22.126][omm2] change /opt/mogdb506/tool/plugins owner to omm2
INFO[2024-06-04T17:41:27.647] [172.20.22.128][omm2] change /opt/mogdb506/tool/plugins owner to omm2
INFO[2024-06-04T17:41:27.665] [172.20.22.128][omm2] installing plugin [pg_prewarm] ...
INFO[2024-06-04T17:41:27.666] [172.20.22.125][omm2] installing plugin [pg_trgm] ...
INFO[2024-06-04T17:41:27.670] [172.20.22.126][omm2] installing plugin [pg_trgm] ...
INFO[2024-06-04T17:41:27.706] [172.20.22.125][omm2] plugin [pg_trgm] installed successfully
INFO[2024-06-04T17:41:27.706] [172.20.22.125][omm2] installing plugin [pg_prewarm] ...
INFO[2024-06-04T17:41:27.717] [172.20.22.126][omm2] plugin [pg_trgm] installed successfully
INFO[2024-06-04T17:41:27.717] [172.20.22.126][omm2] installing plugin [pg_prewarm] ...
......
INFO[2024-06-04T17:41:51.079] finished: exec_post_sql
INFO[2024-06-04T17:41:51.079] begin: restart_dn_finally
INFO[2024-06-04T17:41:51.079] restart all datanodes
INFO[2024-06-04T17:41:51.096] operation: stop
INFO[2024-06-04T17:41:51.096] ========================================
INFO[2024-06-04T17:41:51.160] stop db [172.20.22.128:25000] ...
INFO[2024-06-04T17:41:53.192] stop db [172.20.22.128:25000] successfully
INFO[2024-06-04T17:41:53.192] ========================================
INFO[2024-06-04T17:41:53.192] stop successfully
INFO[2024-06-04T17:41:53.210] operation: stop
INFO[2024-06-04T17:41:53.210] ========================================
INFO[2024-06-04T17:41:53.286] stop db [172.20.22.126:25000] ...
INFO[2024-06-04T17:41:54.318] stop db [172.20.22.126:25000] successfully
INFO[2024-06-04T17:41:54.318] ========================================
INFO[2024-06-04T17:41:54.318] stop successfully
INFO[2024-06-04T17:41:54.334] operation: stop
INFO[2024-06-04T17:41:54.334] ========================================
INFO[2024-06-04T17:41:54.396] stop db [172.20.22.125:25000] ...
INFO[2024-06-04T17:41:56.426] stop db [172.20.22.125:25000] successfully
INFO[2024-06-04T17:41:56.426] ========================================
INFO[2024-06-04T17:41:56.426] stop successfully
INFO[2024-06-04T17:41:56.426] checking cluster state before start
INFO[2024-06-04T17:41:56.480] operation: start
INFO[2024-06-04T17:41:56.480] ========================================
INFO[2024-06-04T17:41:56.480] start db [172.20.22.125:25000] ...
INFO[2024-06-04T17:41:59.384] start db [172.20.22.125:25000] successfully
INFO[2024-06-04T17:41:59.384] start db [172.20.22.126:25000] ...
INFO[2024-06-04T17:42:02.613] start db [172.20.22.126:25000] successfully
INFO[2024-06-04T17:42:02.613] start db [172.20.22.128:25000] ...
INFO[2024-06-04T17:42:05.635] start db [172.20.22.128:25000] successfully
INFO[2024-06-04T17:42:05.836] ========================================
INFO[2024-06-04T17:42:05.836] start cluster successfully
INFO[2024-06-04T17:42:05.836] finished: restart_dn_finally
INFO[2024-06-04T17:42:05.836] upgrade successfully
If you confirm that the upgrade is correct, you can run this command to finish the upgrade:
ptk cluster -n mog506 upgrade-commit
If you want to rollback the upgrade, you can run:
ptk cluster -n mog506 upgrade-rollback [-H|--host HOST] [-i|--id ID]
INFO[2024-06-04T17:42:05.836] time elapsed: 1m41s
[root@mogdb1 ~]# ptk cluster -n mog506 upgrade-commit
INFO[2024-06-04T17:42:13.828] PTK Version: 1.4.7 release
INFO[2024-06-04T17:42:13.833] new step recorder, current step: "restart_dn_finally"
INFO[2024-06-04T17:42:13.833] check cluster status
INFO[2024-06-04T17:42:13.833] checking cluster status
INFO[2024-06-04T17:42:14.106] cluster status is normal
INFO[2024-06-04T17:42:14.106] enable cluster read-write
INFO[2024-06-04T17:42:14.106] set cluster to read-write
INFO[2024-06-04T17:42:14.319] set upgrade_mode to 0
INFO[2024-06-04T17:42:14.350] restore support_extended_features value
INFO[2024-06-04T17:42:14.408] create temp dir
INFO[2024-06-04T17:42:14.423] upload upgrade_extensions.tar.gz
INFO[2024-06-04T17:42:14.441] extract upgrade_extensions.tar.gz to dir /opt/mogdb506/tmp/upgrade
INFO[2024-06-04T17:42:14.619] call hook script: on_commit_script
INFO[2024-06-04T17:42:14.822] clear temporary directories
INFO[2024-06-04T17:42:15.095] commit success
INFO[2024-06-04T17:42:15.095] time elapsed: 1s
[root@mogdb1 ~]#
[root@mogdb1 ~]# ptk cluster status -n mog506
[ Cluster State ]
cluster_name : mog506
cluster_state : Normal
database_version : MogDB-5.0.7 (build c4707384)
mogha_installed : yes
[ Datanode State ]
cluster_name | id | ip | port | user | nodename | db_role | state | uptime | upstream
---------------+------+---------------+-------+------+----------+---------+--------+----------+-----------
mog506 | 6001 | 172.20.22.125 | 25000 | omm2 | dn_6001 | primary | Normal | 00:00:26 | -
| 6002 | 172.20.22.126 | 25000 | omm2 | dn_6002 | standby | Normal | 00:00:23 | -
| 6003 | 172.20.22.128 | 25000 | omm2 | dn_6003 | standby | Normal | 00:00:20 | -
[root@mogdb1 ~]#
用来的真的很丝滑，2分钟就可以完成一套MogDB主备集群的升级。
“免费” 能让生态更活跃
我们认为让更多的人参与进来，能促进生态繁荣，尤其是对我们DBA来讲我认为是天大的好事儿。
为什么这样讲？用的企业多了，如果出问题了，他们找谁？使用量越多，那么DBA的需求就会越旺盛。这个观点我想不用证明了，Oracle、MySQL就是最好的证明。