If you’ve received a message that your lab storage on an RCNFS server (rcnfs01 – rcnfs13, fs2k01, etc.) is moving and you have questions, we hope this page will help answer those. If not, you can reply to the original message or contact us with your question.
Q: What is the process?
A: All members of each lab will be sent an email about scheduling your share’s migration with general information and a link to this FAQ. On the day of the move, your share will be unavailable so that
A) the final sync can complete without any new changes
B) the necessary configuration changes are rolled out across the cluster.
You will be notified when the move is complete and informed of the new full path. Your /n links will remain the same, and any existing Samba/CIFS desktop mount path will remain the same.
Q: Can I still use the lab storage?
A: Yes, up until the day before the move, your storage will be available. A sync from the existing share to the new location has already run and will run again closer to the move. Please refrain from adding or modifying data the day prior to the move if possible. If you add any significant amount of data the day prior, please let us know immediately as an additional sync may be necessary.
Q: Why is this necessary?
A: FASRC is combining physical data center space with the Harvard Medical School in our Boston data center as part of cost-savings measures. We will be working towards moving 18 racks worth of servers to a new location and consolidating as much as possible.
Overall, we will have less rack space and must decommission older, less-efficient servers in favor of more dense systems to make the best use of the space we will have. The RCNFS boxes are very old and are already due for replacement. Your share will be moved to our new CEPH cluster, which is more resilient and not constrained in size as these NFS servers are.
Q: Can it be put off?
A: We must vacate the current space in the data center and will have less rack space in the new area. As this involves work being done by HMS, the Boston data center personnel, and vendors, we must work towards this as quickly as possible. The original schedule was to start in June, but due to the pandemic this was not possible. Each group now has the ability to put limited staff in the data center and we must work against our lease deadline. This migration must be completed by the end of September as we do not have rack space for these servers in the new data center space.
Q: What if we have more than one share on an RCNFS server?
A: We will work with you to consolidate this into one share on the new storage. Most likely this was originally done due to limited contiguous space on the small NFS boxes, which is not an issue with the new storage.
Q: What will happen to your ability to Samba mount your share (if applicable)?
A: We will work with you to maintain your current share’s ability to be mounted if this is the case already. Be aware that you will not be able to mount or access the share via mapping (Samba) on the day of the move.
Q: What will happen to your /n/lab_name link?
A: Your /n symlink will be carried over and point to the new storage. The full path to your storage will change, so if you have the current full path hard-coded in scripts, you will need to update your scripts. Ideally to use the /n/lab_name link instead, which you can change now as that will remain the same.
Q: What is CEPH?
A: CEPH is a dense, distributed storage technology that allows for high resiliency and and can be easily expanded in future. CEPH will feature prominently in our future storage endeavors and is also the foundation the New England Storage Exchange is built upon.
Q: Which share is this?
A: There are around 150, which is why we sent an initial bulk email. Individual emails regarding your share and scheduling details will go out soon. You can also look for your share in this list: