Unable to allocate processing resources

Challenge

Cause

1. Too many jobs are running at a time for the number of concurrent tasks set across all proxies available to the job that failed.
2. Proxies may be unavailable (Shutdown, not connected to the network, etc.).
3. The existing set of backup proxies does not have enough resources to process the backup jobs in a timely manner.
4. The virtual machine might be being processed by another job
5. New LUNs created and/or VMs migrated to new LUNs

Solution

1. Stagger your backup jobs and replication jobs at different time slots so that a job isn’t waiting on available proxies when it starts. (Ex. Instead of scheduling your jobs to all start at 8:00PM, start one job at 8:00, another at 8:30, and another at 9:00.)
2. Increase the number of concurrent tasks allowed on your proxies. However, keep in mind the system requirements.  See How to set max concurrent tasks. If your backup proxy is a virtual machine, you should increase the amount of CPU and RAM available to the proxy.

• VERSION 7.X through to VERSION 9.X with default job settings requires 1 CPU core and 200 MB RAM per task that you wish the proxy to run at a time, plus 2 GB RAM for the operating system and Veeam services. Although only 200 MB per task is required, a 2GB per CPU core ratio is recommended for best performance and reliability. (2 Tasks at once = 2 CPU cores and 2.4 GB RAM required, 4GB recommended)
  • If using high compression, 2 cores per concurrent task are required.
  • Memory usage may be greater than the minimum system requirements. If memory usage on the proxy frequently exceeds 80%, consider increasing the available memory.
  • VERSION 7.X through to 9.X runs tasks in parallel, but can run sequentially with the parallel processing option disabled. Processing multiple disks on one VM will reserve multiple concurrent tasks. See Enabling Parallel Processing.
• Remember to not under-allocate your Veeam Backup server, especially if it is your backup proxy.  If the configuration database is installed on a local SQL instance (the default installation), then SQL Express should be allocated 1 CPU core and 1-1.5 GB RAM when planning your resource usage. A full version of SQL may use significantly more resources in a large environment. 500 MB RAM per concurrent task is recommended for the Veeam Backup services and the SQL server combined; 4 GB minimum.

3. Deploy additional backup proxies from within “Backup Infrastructure->Backup Proxies”.
   • This can be a physical server, an existing VM, or a newly-deployed VM. If adding existing virtual machines, bear in mind that Changed Block tracking will be disabled when backing up these VMs, which can substantially increase processing time.
4. Make sure that your backup job or replication job is selecting the correct proxies. Choose “Automatic selection” to have Veeam try to choose the best proxy from your available pool of proxy machines, or choose a specific set of proxies. The latter can be useful if you have set up a specific proxy with a large amount of resources to be used on a large backup or replication job.

• However, be careful as a backup job set to use a single proxy can fail with the above error if that proxy is unavailable. It might be beneficial to choose more than one if manually specifying proxies.
• Also note that automatic proxy selection will detect available transport modes for each proxy, and then wait for a proxy with the best available transport mode. The modes are ranked as follows: Direct SAN (SAN) > Virtual Appliance (HOTADD) > Network (NBD). For example, if you have a virtual machine proxy on each of three hosts, and all of your VMs are on local storage, a job set to automatic will wait for the proxy that has access to the local storage (HOTADD) rather than use an available proxy that can only read the disk over the network (NBD). For more information, see Transport Modes.

5. Investigate backup job performance. If specific jobs are taking longer to process than normal, check for warnings, compare the bottleneck statistics to previous jobs sessions, and try to isolate the problem to a specific proxy, repository, host, or datastore.
6. Present newly added LUNs to the Veeam Server

More Information