There are many reasons why Archive process failed to archive log file:
1 REDO LOG may be small, so that frequent switch occurs. When ARCH works slowly and LGWR writes fast, this error may appear. Solution is: increase the size of the REDO LOG.
2 If the REDO files are mirrored (duplexed) and they are kept in RAID, then also ARCH process will fail because when one of the REDO Log is filled, Oracle instructs ARCH process to start archive. Immediatley, ARCH process tries to read the 2nd member of the REDO LOG file, which is not yet written (completed) by the DISK controller. Hence, ARCH process fails. To over come, do not keep the redo log files in RAID DISK ARRAYS, keep them non-mirrored (striped ) disk.
3.Avoid both SW level and HW level mirroring of redo log files. One level of mirroring is sufficient for operating a database.
4 If the you operate a VLDB and there are more than 40 disks in the disk array system, keep atleast 6 disks for REDO alone. The rest 34 can be configured for RAID 1+0 or RAID 5.