postgresql - Azure PostgreSQL 突然使用所有存储

Question

昨天，我一直追加的“Azure Database for PostgreSQL 服务器”数据库的增长率突然增加并耗尽了所有存储空间，并且当连接变为只读时，脚本报告了错误。

对磁盘上数据库大小的查询加起来不会达到 18GB 之类的值

    select pg_size_pretty(pg_database_size('database_alpha')) => '80 MB'
    select pg_size_pretty(pg_database_size('database_beta')) => '458 MB'
    select pg_size_pretty(pg_database_size('postgres')) => '8645 kB'

如果不是我的数据库，什么可能正在使用存储以及如何防止将来发生这种情况？

查看日志，下午 5 点左右开始出现很多类似以下内容的条目:

2018-08-26 15:52:39 UTC-5b819a84.45490-WARNING:  archiving write-ahead log file "000000010000004E000000DA" failed too many times, will try again later
2018-08-26 15:52:49 UTC-5b819a84.45490-WARNING:  archiving write-ahead log file "000000010000004E000000DA" failed too many times, will try again later
2018-08-26 15:52:56 UTC-5b819a84.45490-WARNING:  archiving write-ahead log file "000000010000004E000000DA" failed too many times, will try again later
2018-08-26 15:53:00 UTC-5b819a84.45490-WARNING:  archiving write-ahead log file "000000010000004E000000DA" failed too many times, will try again later
2018-08-26 15:53:04 UTC-5b819a84.45490-WARNING:  archiving write-ahead log file "000000010000004E000000DA" failed too many times, will try again later
2018-08-26 15:53:08 UTC-5b819a84.45490-WARNING:  archiving write-ahead log file "000000010000004E000000DA" failed too many times, will try again later
2018-08-26 15:53:11 UTC-5b819a84.45490-WARNING:  archiving write-ahead log file "000000010000004E000000DA" failed too many times, will try again later

Answer 1

我猜测硬件故障阻止了它写入 WAL，并且占用了空间，现在似乎已经恢复了: