A
Status Code 196 will occur when a backup job is in a queued state at the time
the backup window for that job closes. If
a job is active by the time the window closes, that active job will still run
to completion. However, queued jobs will
fail with status 196.
Table of Contents
1 Managing
Backup Policies and Storage Units
Effective policy
management plays a major roll in completing backups within the time
window. The first thing to consider is
the backup window itself. Determine if
the current backup window is reasonable for the volume of data being backed up
and the available storage resources.
1.1
Policy Configuration
1.1.1
Maximum Jobs per policy and client
Policies have a
setting for “Limit jobs per policy”. Ensure
that this setting is not overly constrictive of jobs that could be
running. If the maximum number of jobs
is reached on the policy, additional jobs will queue. If this setting is too low, queued jobs may
result in a status 196 despite the fact that they have the resources to run.
Within “Global
Attributes” under the Master Server’s host properties, there is a setting for
the “Maximum jobs per client”. This has
an effect similar to that of “Limit jobs per policy”, but on the client level. If multiple data streams are being used,
ensure that this setting is not constricting jobs that could be running. If the maximum number of jobs is reached on
the client, additional jobs will queue.
1.1.2
Additional Features/Options
Please take note of
the following features, as they are designed to use the backup window more efficiently:
·
Synthetic backups:
Synthetic backups allow logical full and
cumulative “backups” to take place without actually transferring data from the
client. This can improve performance by providing
the restore benefits of Full or Cumulative Incremental backups, without backing
up redundant data during the backup window.
·
Advanced Client
backups: Advanced client
backup methods such as Flashbackup involve “snapshot” backups. This can improve performance by backing up
the “raw” data, which eliminates the overhead of querying every file on the
file system level.
·
Disk staging: Disk
staging allows the “staging” of data to disk before ultimately moving to the
final storage unit. This can reduce the
required backup window by deferring the move from disk to tape until all client
data is collected and the backup window closes.
2
Optimizing Backup Performance
Another method of avoiding the 196 error is
to improve overall backup speed in order to complete all jobs within the
designated window. This section offers
some general guidelines for optimizing performance.
2.1 Identify
Bottlenecks
Identifying where
bottlenecks in the data flow exist is essential to improving performance. The following are some common methods
of doing so:
1. Are there any common
factors between poorly performing clients, such as OS platform, Network, or the
Media Server being used?
2. Perform a backup to a
disk storage unit for a poorly performing client. Is performance significantly improved? This can eliminate a tape storage unit and
its data path as being part of the problem.
3. Determine whether the
media server experiences performance issues during a local backup (backing up
locally attached storage). This can eliminate
the network and client as the primary bottlenecks.
4. Perform a backup to
“null” on a poorly performing client.
This allows the NetBackup client to process the data without actually
sending it to the media server for backup.
This can isolate the client itself as the bottleneck. See NetBackup technote 242918 for information
on executing and analyzing this test.
The
results of tests such as these can assist in identifying areas to investigate
further. Dependant on the results, some
next steps may include:
- Checking for optimum Network
duplex settings.
- Verifying the OEM recommended
driver and firmware versions are installed for tape hardware.
- Adding additional hardware to
satisfy the backup requirements (tape devices, SCSI / fiber equipment,
etc.)
- Enabling the NetBackup
multiplexing feature.
- Adjusting tape buffer sizes.
2.2 Multiplexing
Multiplexing
is a NetBackup feature allowing multiple streams of data to be written to a
tape device concurrently. These streams
may come from a variety of clients, even of dissimilar platforms.
At
times, a tape device may be capable of writing much faster than a single client
can present data to the media server. In
this scenario, multiplexing can significantly improve overall performance.
Implementation of multiplexing should always
be carefully considered. An environment
which unnecessarily uses multiplexing may see no benefits or even diminished
backup performance, and will always see a negative impact on restore performance.
See the Multiplexing section of your
NetBackup System Administrator’s Guide (reference Multiplexing in the index) for
a full explanation on the multiplexing feature.
2.2.1 Multiplexing
- Planning
Use
the results from section 2.1 (Identifying Bottlenecks) to help determine if
multiplexing is appropriate. For
example, if it is found that there are no problems within the Media Server or
the network, but the client processing is very slow, multiplexing may help significantly.
Testing
of both backup and restore performance should be done to determine the ideal
multiplexing level for an environment.
As an example, a value of three (at the storage unit level) is often
found to be a good trade off between backup and restore performance.
2.2.2 Multiplexing
- Implementation
Multiplexing is configured in three places:
·
Within
the Storage Unit, under “Maximum multiplexing per drive”.
·
Within
the Policy Attributes through the “Allow Multiple Data Streams” checkbox.
·
Within
the Policy’s Schedule in the “Media Multiplexing” setting.
As mentioned above, the following
configuration options might also be relevant when configuring multiplexing:
·
Limit
jobs per policy (set in Policy Attributes)
·
Maximum
Jobs per Client (set in Global Attributes)
·
Maximum
Data Streams (set in ‘Client Attributes’ within the Master Server’s host
properties)
See the VERITAS NetBackup System
Administrator’s Guide for further information on multiplexing configuration.
2.3 Eliminate
Unnecessary Logging
When not actively troubleshooting a NetBackup
issue, debug logs should be disabled or kept at a very low verbose level. NetBackup debug logs can grow extremely
large, and consume gigabytes of space under certain configurations. Verbose logging can also have a negative
impact on performance. The bpdbm log particularly has a
significant impact on resources, even when at a verbose level of 0.
See the NetBackup System Administrators guide
for information on creating and disabling logging, as well as manipulating the
verbose level.
3 Verify
hardware availability
Hardware problems
(tape drive or robotic problems) can result in down tape drives. This means that NetBackup has fewer resources
with which to complete backups. Use the
NetBackup Device Monitor GUI to view the status of all the tape drives. If there are any down tape drives, examine
the operating system’s error logs to determine why the drives were “down’ed”. Running
the Problems Report in the NetBackup Administration console may also assist in
diagnosis. Additionally, it may be
necessary to reconfigure, or verify the configuration of devices. See the Media Manager and Device
Configuration Guides for more information on how to proceed with this.
No comments:
Post a Comment
Note: only a member of this blog may post a comment.