Instance Architecture   «Prev 

Archiver (ARC0)

I examined my alert log and noticed that my archive process is failing with the ARC0: Failed to archive log message.
Question: Can someone suggest how this alert log error can be corrected?
ARC0: Beginning to archive log# 2 seq# 313533
ARC0: Failed to archive log# 2 seq# 313533
Thur May 07 07:42:57 2015
ARC1: Failed to archive log# 3 seq# 313532
Thur May 07 07:42:57 2015
ARC1: Beginning to archive log# 2 seq# 313533
ARC1: Failed to archive log# 2 seq# 313533
Thur May 07 07:43:31 2015
ARC0: Beginning to archive log# 3 seq# 313532
ARC0: Failed to archive log# 3 seq# 313532
ARC0: Beginning to archive log# 2 seq# 313533
ARC0: Failed to archive log# 2 seq# 313533
Answer: This alert log message "failed to archive log" is not an "error", it is just an informational message noting that there was a delay when archiving your redo log, usually due to a disk I/O slowdown.
You have not lost any redo data, but this error does indicate that you have a redo-related bottleneck, usually disk I/O.
  1. Make sure that your redo is on a separate disk spindle
  2. If necessary, move your redo to faster disk (SSD) devices.
It seems like your logfiles are not keeping up write activity and you should either
  1. move your logfiles to speedy media to avoid write contention, or
  2. try adding another log group members or resize redo log files with bigger number.

As changes are made to an Oracle database, a log of those changes is written to the redo log files.

As each redo log is filled, the archiver process will begin to copy it to the archive.

The archiver will more or less keep up with the log writer.

If the log writer gets far enough ahead to fill up all the redo log files, then database users will be forced to wait while the archiver copies another file.

When a database is running in archive log mode, then log writer will not overwrite a log file until it has been copied by the archiver.

Once a log file has been copied, processing will resume. The archived log files preserve a record of all changes and is used for disaster recovery.