Dear CRCD Users,
As a follow-up to our previous alert, the worst of the performance problems on the /ix and /ix1 storage filesystems have subsided, and interactive response across the clusters should be noticeably better. We will continue to monitor the filesystems.
As a reminder, please run I/O-intensive and AI workloads from /vast or local scratch rather than /ix and /ix1. Adhering to this best practice will prevent the same disruptions from coming back and impacting you and others using the system.
If you continue to experience performance issues, please submit a help ticket.
- The CRCD Team