Preemption is defined by the Job Manager at each phase of a job. Jobs that can be interrupted by the Job Manager or by the user and then restarted without having to start the phase from the beginning are called Preemptable jobs. A non-preemptable job is one that cannot be interrupted by the Job Manager or suspended by the user.
If a running job is preemptable, the Job Manager can interrupt the running job and allocate the resources to a higher-priority job if the needed resources are streams, drives or media. The interrupted job enters a waiting state and resumes when the resources it needs become available. Backup and restore operations preempt auxiliary copy and other jobs (except backup). The preempting option is enabled by default, to disable this feature, see configuring restartability for a job type.
The following table provides information on the Status of the job in the Job Controller window and the Reason for job delay displayed in the Job Details dialog box when a job is preempted. In addition, a brief explanation on what happens when a job is preempted is also provided.
Jobs |
Status in the Job Controller |
Reason for Job Delay |
Additional Information |
---|---|---|---|
Backup Operation |
|
|
Once interrupted, job does not hold on to resources and returns to Waiting status. The job retries for resources. (The Status of the job in the Job Controller window and messages in the Reason for job delay are discussed in What Happens When There are no Resources for a Job.) |
Data Recovery Operations (for File System-like agents) |
|
|
Once interrupted, job does not hold on to resources and returns to Waiting status. The job retries for resources. (The Status of the job in the Job Controller window and messages in the Reason for job delay are discussed in What Happens When There are no Resources for a Job.) |
Data Recovery Operation (for Database-like agents) |
Not Preemptable |
Not Preemptable |
Not Preemptable |
Index Restore (Browse Backup Data) |
Not Preemptable |
Not Preemptable |
Not Preemptable |
Auxiliary Copy |
|
|
Once interrupted, job does not hold on to resources and returns to Waiting status. The job retries for resources. (The Status of the job in the Job Controller window and messages in the Reason for job delay are discussed in What Happens When There are no Resources for a Job.) |
Synthetic Full |
|
|
Once interrupted, job does not hold on to resources and returns to Waiting status. The job retries for resources. (The Status of the job in the Job Controller window and messages in the Reason for job delay are discussed in What Happens When There are no Resources for a Job.) |
Media Refresh |
Waiting |
No resources available |
Once interrupted, job does not hold on to resources and returns to Waiting status. The job retries for resources. |
The higher priority job that is doing the resource preemption will display the Reason for Job delay as follows:
Waiting for job[ ] to release the resources.
Preemptable and Non-Preemptable Jobs
In a preemptable phase, the job can be interrupted by the Job Manager or suspended by the user and then restarted without having to start the phase over again from the beginning. Preemptable jobs are always restartable. A File System backup phase is one example of a preemptable phase; the Job Manager can interrupt this phase when resource contention occurs with a higher priority job.
A non-preemptable phase is one that cannot be interrupted by the Job Manager or suspended by the user. It can only run to completion, be killed by administrative action, or be failed by the system. For example, the data recovery operations of database agents are non-preemptable.
The following table lists the types of preemptable and non-preemptable jobs:
Preemptable and Restartable |
Non-preemptable and Non-Restartable |
Non-preemptable but Restartable |
---|---|---|
Data protection operations for most non-database agents. For more information, see Supported and Unsupported Agents for Preemptable Backups. |
Data recovery operations for database-like agents. |
Data protection operations for database agents. |
DataArchiver archive jobs during the Archive Index and Archive Content Index phases of the job. |
Media export, erase media, and inventory jobs. |
The system state phase of Windows File System data protection operations. |
Data recovery operations for most File System-like (indexing-based) agents during the restore phase. |
SAN volume data protection jobs (non-preemptable in its scan phase). |
Offline Content Indexing jobs. |
Data recovery operations from the Search Console. |
Disk volume reconciliation jobs. |
|
Most administration jobs including Install Automatic Updates and Download Automatic Updates. |
||
Silo backup and restore operations. |
||
Media refresh operations. |
||
Deduplication database reconstruction job. |
Supported and Unsupported Agents for Preemptable Backups
The following is a list of agents that support preemptable backups.
-
Sybase Database
-
Unix Notes Database
-
Notes Document
-
NDMP
-
HP-UX File System
-
HP-UX 64bit File System
-
Solaris File System
-
Solaris 64bit File System
-
AIX File System
-
Oracle Database
-
ProxyHost
-
Linux File System
-
Windows File System
-
Exchange Public Folder
-
Unix Tru64 64-bit File System
-
DB2
-
Proxy Client File System
-
ProxyHost on Unix
-
Active Directory
-
Notes Database
-
Unix Notes Document
-
Exchange Database
-
Exchange Mailbox (Classic)
-
Exchange Web Folder
-
Exchange Mailbox Archiver
-
Windows File Archiver
-
Notes Database
-
Notes Document
-
SAP for Oracle
-
MAC FileSystem
-
Big Data Apps
-
Image Level
-
Unix File Archiver
-
Exchange Compliance Archiver
-
File Share Archiver
-
FreeBSD
-
Image Level On Unix
-
Image Level ProxyHost
-
Unix Notes Database
-
SharePoint Server
-
SAP for MAX DB
-
Oracle RAC
-
SQL Server
-
Exchange PF Archiver
-
Continuous Data Replicator
-
OES File System on Linux
-
Image Level ProxyHost on Unix
-
MS SharePoint Archiver
-
Domino Mailbox Archiver
-
DPM
-
DB2 MultiNode
-
MySQL
-
Virtual Server
-
Object Link
-
PostgreSQL
-
External Data Connector
-
Documentum
-
Object Store
-
Cloud Apps
-
SAP HANA
-
Exchange Mailbox
-
Azure AD
-
Azure Web App
The following is a list of agents that do not support preemptable backups.
-
Informix IDA
-
OSSV Agent
-
Groupwise DB on OES
-
Recovery Director
For a list of VSA agents that support preemptable backups, see Virtualization.
What Happens When There Are No Resources for a Job
Each job requires certain resources for its successful completion. Absence of these resources affects different types of jobs differently. The following table discusses the resources required by each job, the status of the job in the Job Controller window when there are no resources and the corresponding examples of the Reason for job delay displayed in the Job Details dialog box. In addition, a brief explanation on what happens when a job does not have the required resources is also provided.
By default the Commvault Media & Library Manager service on the CommServe cleans up any media and drive reservation that is held by a job which failed to release the resource when it was abruptly terminated, every 1440 minutes. You can modify the frequency using the nRESOURCERELEASEINTERVALMIN registry key.
Jobs |
Resources |
Status in the Job Controller |
Reason for Job Delay |
Additional Information |
---|---|---|---|---|
Backup Operation |
Streams, Active Media, Drive |
Waiting Waiting |
See Example 1. See Example 2. |
Job checks for necessary resources. If the resources are not available the job retries to reserve the resources when ever they are freed. Does not hold on to any resource until all the necessary resources are available. |
Data Recovery Operations (for File System-like agents) |
Drive |
Pending |
The media is already reserved by some other job(s). |
If the resources are not available the job retries to reserve the resources when ever they are freed. |
Data Recovery Operation (for Database-like agents) |
Drive |
Failed Running |
See Example 1. See Example 2. |
Job checks for necessary resources. If the resources are not available it retries every 2 minutes to reserve the resources. Does not hold on to any resource until all the necessary resources are available. |
Index Restore Operation (Browse Backup Data) |
Destination Drives Source Media |
Pending Waiting Waiting Running Pending |
See Example 1. See Example 2. See Example 2. |
Job checks for necessary resources. Job reserves 2 drives for source and destination media. If the above resources are not available, it retries every 2 minutes to reserve these resources. Does not hold on to any resource until all the necessary resources are available. Once the 2 drives and destination media is obtained job reserves the source media. If the job encounters resource contention while reserving the source media, (Example 2) it retries every 20 minutes and a maximum of 144 times to obtain the source media. Holds on to the 2 drives and destination media as long as it is not interrupted and as long as the source media is available. |
Synthetic Full |
Streams, Destination Drives, Destination Media Source Media |
Waiting Waiting Waiting Running Pending |
See Example 1. See Example 2. See Example 2. |
Job checks for necessary resource Job reserves streams, marks active media full, reserves 2 drives and destination media If the resources are not available the job retries to reserve the resources whenever they are freed. Does not hold on to any resource until all the necessary resources are available. Once the 2 drives and destination media is obtained job reserves the source media. If the job encounters resource contention while reserving the source media, (Example 2) it retries every20 minutes and a maximum of 144 times to obtain the source media. Holds on to the 2 drives and destination media as long as it is not interrupted. |
Example 1
The operation could not be completed as the drive pool is offline.
The operation cannot be completed as the host is offline.
The operation could not be completed as the library controller is offline.
The operation cannot be completed as the library is offline.
The operation cannot be completed as the master drive pool is offline.
There are not enough drives in the drive pool.
There is no active controller for this library.
Mount path is offline.
Media agent is offline.
Host is offline.
Library controller service is down.
The Library controller is offline.
Example 2
All spares are outside the library.
The operation cannot be completed as the drive is reserved.
The operation cannot be completed as the media is outside the library.
The operation cannot be completed as the mountpath is reserved.
The operation could not be completed as there is not enough media.
The operation could not be completed as there are not enough drives that are not reserved, online and whose controller are alive.
There is no active controller for this drive.
There are no disk paths that are enabled and with the required space.
There are no more spares.
The operation cannot be completed as the opposite side for this media is already reserved.
The operation cannot be completed as the number of drive reservations exceeds the allotted drives.
The number of writers would exceed the maximum allowed limit.
The operation cannot be completed as the media group is reserved.
The operation cannot be completed as the copy is reserved.
The operation cannot be completed as the drive pool is used by storage policy copy.
Requested volume is being mounted/unmounted.
The media is not in library.
Media is being used and is reserved.
Job does not have reservation on the drive.
The requested media is not in any slot of the library.
The requested media is stuck in the drive.
The requested media is exported.
The media is already Reserved by some other Job[s].
The Media is not available. The Job currently using the media was interrupted.
No drives available for reservation.
Not enough good drives available for reservation.
Not enough drives available in Drive Pool.
The interrupted job has not released the drives yet.
The interrupted job has not released the Media yet.
The job has already been interrupted by another job.
Job[s] Interrupted by this Job have not released Resources yet.
Not Enough streams Available for Storage policy [^1%s]. Need ^2%d stream[s] and ^3%d stream[s] are available.
The Media is already reserved by some other Job[s].
Waiting for Jobs [ ] to release the resources.
No resources available.