Linux, FreeBSD, Juniper, Cisco / Network security articles and troubleshooting guides

How to copy files between two SRX cluster nodes - remote procedure

A Juniper SRX chassis cluster is formed by joining two (max) devices that act as a single chassis. Some of the enterprise networks allow management of both nodes via the management port fxp0 management interface. But sometimes, the environment allows managing the cluster via the revenue port, specifically the redundant (reth) ethernet interface.

The trick with these reth interfaces is that they are formed of members physical fe/ge/xe interfaces of both nodes and each reth is assigned to only one redundancy group in the cluster. Each redundancy group is active only on one node at a time. So far this is not an issue (specially if there are two redundancy groups for the data plane - RG1 and RG2).

The inconvenience comes when there are files that need to be copied to/from both nodes of the cluster: request support information (RSI) files, archived logs or something else. Some would convenient to move RG0 (the routing engine) to each node individually and copy the necessary files via the reth interface. With only static routes this is not a problem. When dynamic routing protocols are configured (without graceful restart).

The more interesting approach is to use the cli “>file copy” operational command or “rcp” unix shell command. Let’s look at both methods.

Transferring files between SRX cluster nodes with file copy.

Transferring files between SRX cluster nodes with rcp.

Both of the above commands copy files to the other cluster node, they are being pushed. In order to copy files from the other node, the arguments have to be reversed.