Veritas NetBackup™ Deduplication Guide

Media write error (84) if due to a full local cache directory

One possible cause for a media write error (84) in a CloudCatalyst environment is described in the following topic.

The administrator configures a local cache directory as part of configuring a CloudCatalyst storage server. Cache eviction is triggered when data causes this directory to reach the high watermark setting (HighWatermark). Once the high watermark is reached, data is purged when the used space reaches the midpoint between HighWatermark and LowWatermark (high+low)/2 and continues until LowWatermark is reached. If the rate of incoming data exceeds the rate where the watermark can be maintained, the jobs begin to fail. In this situation, the directory does not have enough space for new files and the MSDP write operation fails with a media write error (84).

This error is more likely to occur when a relatively small disk is used for demonstration or proof-of-concept testing. To prevent this error in a production environment, make sure that there is at least 4 TB of disk space available for the cache. Make sure that the cache is located on a dedicated file system that is used only for the CloudCatalyst local cache.

After receiving a media write error (84) and addressing the cause, restart all NetBackup services to avoid a media open error (83).

The UNIX df command can be used to determine the disk space in the local cache directory. For example:

df /esfs
Filesystem     1K-blocks  Used       Available  Use%  Mounted on
/dev/loop0     1007896    1007896    0          100%  /esfs

See the following topic for more information about CloudCatalyst local cache considerations:

See About the CloudCatalyst cache.