上一篇我们学习了日志清理,日志清理虽然解决了日志膨胀的问题,但就无法再恢复检查点之前的一致性状态。因此,我们还需要日志归档,pg的日志归档原理和Oracle类似,不过归档命令需要自己配置。
pg主要归档参数如下:
archive_mode参数有3种模式:
以下代码在postmaster.c
- /*
- * Archiver is allowed to start up at the current postmaster state?
- *
- * If WAL archiving is enabled always, we are allowed to start archiver
- * even during recovery.
- */
- #define PgArchStartupAllowed() \
- (((XLogArchivingActive() && pmState == PM_RUN) || \
- (XLogArchivingAlways() && \
- (pmState == PM_RECOVERY || pmState == PM_HOT_STANDBY))) && \
- PgArchCanRestart())
除了开启归档外,还需要保证wal_level不能是MINIMAL状态(因为该状态下有些操作不会记录日志)。在db启动时,会同时检查archive_mode和wal_level。以下代码也在postmaster.c(PostmasterMain函数)。
- if (XLogArchiveMode > ARCHIVE_MODE_OFF && wal_level == WAL_LEVEL_MINIMAL)
- ereport(ERROR,
- (errmsg("WAL archival cannot be enabled when wal_level is \"minimal\"")));
pg会启动一个辅助进程,作用是实时监控事务日志,发现能归档的日志则会通过用户设置的archive_command参数中的命令进行归档。归档命令可以很自由地被指定,一般是cp或者加上压缩命令,如果设置该参数,或者命令有错误,则无法真正归档。
以下代码在pgarch.c(pgarch_ArchiverCopyLoop函数)
- /* can't do anything if no command ... */
- if (!XLogArchiveCommandSet())
- {
- ereport(WARNING,
- (errmsg("archive_mode enabled, yet archive_command is not set")));
- return;
- }
如果只在日志切换时归档,假如在日志段未满时宕机,则归档日志会缺失一部分,可能造成数据丢失。另外,如果业务写请求较少,日志可能长期不归档。此时,可以通过archive_timeout参数设置超时强制归档,提高归档频率。
注意,每次日志切换时,即使未写满日志大小依然是16M,因此该参数如果设置太小,可能导致归档过于频繁并且大量浪费空间。
代码在checkpointer.c文件(CheckArchiveTimeout函数),有一丢丢长,我们放在下面看。
每当WAL日志段切换时,就可以通知日志归档进程将该日志进行归档。
日志写入函数(一个老熟人),当一个日志段写满时,需要切换。
- static void
- XLogWrite(XLogwrtRqst WriteRqst, bool flexible)
- {
- …
- /* 一个段已满 */
- if (finishing_seg)
- {
- /* 将该段刷入磁盘,保证归档日志数据完整性 */
- issue_xlog_fsync(openLogFile, openLogSegNo);
-
- /* 通知WalSender进程发送日志给从库 */
- WalSndWakeupRequest();
-
- LogwrtResult.Flush = LogwrtResult.Write; /* end of page */
-
- /* 发送日志归档的通知信息 */
- if (XLogArchivingActive())
- XLogArchiveNotifySeg(openLogSegNo);
- …
- }
- }
创建归档通知的.ready文件(相当于一种进程间的通信机制),告诉归档进程应该归档哪个日志。
- /*
- * Create an archive notification file
- *
- * The name of the notification file is the message that will be picked up
- * by the archiver, e.g. we write 0000000100000001000000C6.ready
- * and the archiver then knows to archive XLOGDIR/0000000100000001000000C6,
- * then when complete, rename it to 0000000100000001000000C6.done
- */
- void
- XLogArchiveNotify(const char *xlog)
- {
- char archiveStatusPath[MAXPGPATH];
- FILE *fd;
-
- /* insert an otherwise empty file called
.ready */ - StatusFilePath(archiveStatusPath, xlog, ".ready");
- fd = AllocateFile(archiveStatusPath, "w");
- if (fd == NULL)
- {
- ereport(LOG,
- (errcode_for_file_access(),
- errmsg("could not create archive status file \"%s\": %m",
- archiveStatusPath)));
- return;
- }
- if (FreeFile(fd))
- {
- ereport(LOG,
- (errcode_for_file_access(),
- errmsg("could not write archive status file \"%s\": %m",
- archiveStatusPath)));
- return;
- }
-
- /* Notify archiver that it's got something to do,发送信号给日志归档进程(旧版本是先发给Postmaster进程) */
- if (IsUnderPostmaster)
- PgArchWakeup();
- }
发送信号通知归档进程
- /*
- * Wake up the archiver
- */
- void
- PgArchWakeup(void)
- {
- int arch_pgprocno = PgArch->pgprocno;
-
- /*
- * We don't acquire ProcArrayLock here. It's actually fine because
- * procLatch isn't ever freed, so we just can potentially set the wrong
- * process' (or no process') latch. Even in that case the archiver will
- * be relaunched shortly and will start archiving.
- */
- if (arch_pgprocno != INVALID_PGPROCNO)
- SetLatch(&ProcGlobal->allProcs[arch_pgprocno].procLatch);
- }
实际上日志归档的顺序也很重要,归档进程会优先选择段号较小的日志文件。因为日志清理时也是按段号顺序清理的,段号小的日志优先归档完就可以被清理了。
归档完成后,归档进程会将.ready文件改为.done文件。
- /*
- * pgarch_ArchiverCopyLoop
- *
- * Archives all outstanding xlogs then returns
- */
- static void
- pgarch_ArchiverCopyLoop(void)
- {
- char xlog[MAX_XFN_CHARS + 1];
-
- /*
- * 循环处理.ready文件
- */
- while (pgarch_readyXlog(xlog))
- {
- int failures = 0;
- int failures_orphan = 0;
-
- for (;;)
- {
- struct stat stat_buf;
- char pathname[MAXPGPATH];
-
- /*
- * 如果收到停库请求或者postmaster异常挂掉,不再执行后续操作,直接返回
- */
- if (ShutdownRequestPending || !PostmasterIsAlive())
- return;
-
- /*
- * Check for barrier events and config update. This is so that
- * we'll adopt a new setting for archive_command as soon as
- * possible, even if there is a backlog of files to be archived.
- */
- HandlePgArchInterrupts();
-
- /* 如果没有设置archive_command或者设置有问题,报错返回 */
- if (!XLogArchiveCommandSet())
- {
- ereport(WARNING,
- (errmsg("archive_mode enabled, yet archive_command is not set")));
- return;
- }
-
- /* 一段异常宕机导致出现孤儿.ready文件时的处理,略 */
- …
-
- /* 进行日志归档 */
- if (pgarch_archiveXlog(xlog))
- {
- /* successful,归档成功,将.ready改为.done文件 */
- pgarch_archiveDone(xlog);
-
- /*
- * Tell the collector about the WAL file that we successfully archived
- */
- pgstat_send_archiver(xlog, false);
-
- /* 开始处理下一个日志 */
- break; /* out of inner retry loop */
- }
- /* 归档失败 */
- else
- {
- /*
- * Tell the collector about the WAL file that we failed to
- * archive
- */
- pgstat_send_archiver(xlog, true);
- /* 如果失败次数大于重试次数,报错返回 */
- if (++failures >= NUM_ARCHIVE_RETRIES)
- {
- ereport(WARNING,
- (errmsg("archiving write-ahead log file \"%s\" failed too many times, will try again later",
- xlog)));
- return; /* give up archiving for now */
- }
- pg_usleep(1000000L); /* wait a bit before retrying,休眠1秒,重试 */
- }
- }
- }
- }
- /*
- * CheckArchiveTimeout -- check for archive_timeout and switch xlog files
- */
- static void
- CheckArchiveTimeout(void)
- {
- pg_time_t now;
- pg_time_t last_time;
- XLogRecPtr last_switch_lsn;
-
- /* 未设置超时参数或者在恢复阶段,直接返回 */
- if (XLogArchiveTimeout <= 0 || RecoveryInProgress())
- return;
-
- now = (pg_time_t) time(NULL);
-
- /* First we do a quick check using possibly-stale local state.
- 首先快速检查,看当前时间减去本地保存的last_xlog_switch_time是否超时,没有则返回
- */
- if ((int) (now - last_xlog_switch_time) < XLogArchiveTimeout)
- return;
-
- /*
- * Update local state ... note that last_xlog_switch_time is the last time a switch was performed *or requested*.
- 从共享内存中获得上次日志切换的时间,这是真正的日志切换时间。同时获取上次日志切换的LSN
- */
- last_time = GetLastSegSwitchData(&last_switch_lsn);
-
- /* 取两者较新的时间,更新本地保存值 */
- last_xlog_switch_time = Max(last_xlog_switch_time, last_time);
-
- /* Now we can do the real checks,真正的检查,如果超时,执行后面的检查 */
- if ((int) (now - last_xlog_switch_time) >= XLogArchiveTimeout)
- {
- /*
- * Switch segment only when "important" WAL has been logged since the
- * last segment switch (last_switch_lsn points to end of segment
- * switch occurred in).
- 如果日志的“重要”LSN >上次切换的LSN,则说明自上次切换以来有重要的WAL日志写入,执行强制切换日志段
- */
- if (GetLastImportantRecPtr() > last_switch_lsn)
- {
- XLogRecPtr switchpoint;
-
- /* mark switch as unimportant, avoids triggering checkpoints,切换日志段 */
- switchpoint = RequestXLogSwitch(true);
-
- /*
- * If the returned pointer points exactly to a segment boundary,
- * assume nothing happened. 如果返回的指针正好在段边界,当做无事发生。否则记录一条DEBUG1级别的切换信息
- */
- if (XLogSegmentOffset(switchpoint, wal_segment_size) != 0)
- elog(DEBUG1, "write-ahead log switch forced (archive_timeout=%d)",
- XLogArchiveTimeout);
- }
-
- /*
- * Update state in any case, so we don't retry constantly when the
- * system is idle. 更新切换时间
- */
- last_xlog_switch_time = now;
- }
- }
参考
《PostgreSQL技术内幕:事务处理深度探索》第4章