Hawk Maintenance 1st July 2020

We are scheduling an at risk/partial maintenance period for the Hawk cluster next Wednesday (July 1st) 0900-1800.

At least three potentially disruptive (to varying degrees) activities will be taking place:

i) system security enhancements

ii) remounting of external storage and maintenance of scratch re-exporter (affects webmo/smartlink)

iii) security penetration testing of the Hawk system (more details bottom)

Security enhancements and storage works:

-All users who login using SSH keys will need to setup their ‘.ssh/authorized_keys’ file again after the maintenance period as we will be performing a system-wide cleanup.  This annual activity is being undertaken to improve security of the system.

-The my.supercomputing.wales (MySCW) profile pages may also be unavailable at times during the maintenance period

-During this time please expect interruptions to external storage connections  (e.g. gluster) and there will be interruptions to the scratch re-export (serving webmo/smartlink). There will also be  periods where users will not be allowed to login to the login nodes.

-Please ensure that you aren’t running jobs or activities requiring gluster storage on Wednesday July 1stWebmo jobs or smartlink jobs will also be unable to run during this time as the scratch re-exporter is worked upon.

-Jobs which utilise standard hawk nodes and native scratch and home storage should continue to run unaffected.

Security penetration testing details:

Throughout the period of 1st July – 17th July, we are planning to undertake a full penetration test of the Hawk system. 

We will be scheduling Wednesday 1st July as a regular maintenance window for Hawk. We will schedule any penetration test work that has a potential to disrupt service during this period. We have been advised by our suppliers that this will only likely amount to an increased network load on the system, which may result in delayed responsiveness from our servers. However as there is a small chance that the work may impact on a greater scale, the system should be considered at-risk.

We do not expect that the bulk of the work will impact on any services, and expect that you will see no difference in the normal operation of the system during this period outside of the maintenance window on the 1st July.

Leave a Reply

Your email address will not be published.