Requirements

Overview

Note: Before choosing your deployment profile, see Supported use cases for single-node and multi-node installations.

What you need	Requirements	Configuration	Installation
One Linux machine (RHEL 8.2, 8.3, 8.4, 8.5)	Multi-node machine requirements	Configuring the machines	Online multi-node installation Offline multi-node installation
SQL Server (MS SQL 2016, 2017, 2019 - Enterprise and Standard)	-	Configuring MS SQL Server	-
DNS	-	Configuring the DNS	-
Load balancer	-	Configuring the load balancer	-

Multi-node Machine Requirements

This is the only configuration supported for production deployments.

Bundle	Minimum	Recommended
Node count	At least 3 server nodes. There must be an odd number of server nodes in a cluster to have increased fault tolerance Any number of agent nodes.	At least 3 server nodes. There must be an odd number of server nodes in a cluster to have increased fault tolerance Any number of agent nodes.
(v-)CPU total	48 (v-)CPU	64 (v-)CPU
Minimum (v-)CPU per node	12 (v-)CPU	16 (v-)CPU
RAM total	96 GiB	144 GiB
Cluster disk for each node	256 GiB SSD Min IOPS: 1100	256 GiB SSD Min IOPS: 1100
Data disk for each server node	2 TiB SSD Min IOPS 1100	2 TiB SSD Min IOPS 1100
etcd disk for each server node	16 GiB SSD Min IOPS: 240	16 GiB SSD Min IOPS: 240
UiPath Bundle Disk (For offline installation only, on the first server node)	512 GiB SSD Min IOPS: 1100	512 GiB SSD Min IOPS: 1100

Note: Following the minimum requirements above is enough for three HA skills (two replicas) and one pipeline running in parallel with the default configuration.

Additional agent node with GPU support for Document Understanding

Bundle	Requirement
(v-)CPU	8 (v-)CPU
RAM	52 GiB
Cluster disk	256 GiB SSD Min IOPS: 1100
Data disk	N/A
GPU RAM	11 GiB

Advanced storage and compute requirements

Use	CPU	RAM (GiB)	GPU
Hardware usage right after install	30 (for minimum configuration) 33 (for recommended configuration)	48	0
Minimum for serving (ML Skill, one replica)	0.6	2	0
Minimum for Training (Pipeline)	1	4	0
Document Understanding model Serving (ML Skill, one replica)	1	4	0
Document Understanding model Training	2	24	1

Note: Trainable Document Understanding ML Packages provided by UiPath will work on both CPU or GPU. For 2021.10, training on GPU is at least 5 times slower than training on CPU. Also, please know that training on CPU is only supported for datasets up to 5000 pages in size. Validation Station retraining loop is not supported on deployments without a GPU, due to the fact that dataset sizes increase too fast and may hit a compute wall with CPU very quickly. If you run UiPath Document OCR Local Server on AI Center to process more than 2 million pages of documents a year, GPU is strongly recommended for a better product experience.