Workers with 100G SmartNICs in maintenance mode on multiple sites

[Ilya Baldin]

Hello,

We have placed worker nodes with 100G ConnectX-6 SmartNICs into maintenance mode at the following sites:

• CLEM

• DALL

• FIU

• GATECH

• MASS

• MICH

• NCSA

• STAR

• TACC

This is based on the need to flash FPGAs in those worker nodes this week in preparation for a seminar being taught on FABRIC. Once the activity is completed the nodes will be returned in service. We expect most reservations on those nodes will expire during maintenance, for those that do not we will do our best to return the VMs back in service after we reboot the underlying servers (this is a requirement for FPGA flashing).

The rest of the infrastructure at those sites is unaffected – servers with GPUs as well as servers with ConnectX-5 10/25Gbps SmartNICs are still available.

[Mert Cevik]

Dear Experimenters,

We completed this maintenance, worker nodes mentioned on the sites listed on the previous message are available for experiments.
None of the VMs or reservations running on the workers with 100G ConnectX-6 SmartNICs were affected during the work.

[Author]

Posts