The /home file system

The /home file system is the main area where our users keep their data. Each user's home directory resides there. It is called /global/home/hpcXXXX where hpcXXXX denotes the user name. This system is shared, i.e. it is visible from all our servers and the workup/login nodes.

Physically, this file system resides in 4 racks containing approximately 1 PB of raw disk space. The configuration is designed to tolerate the failure of multiple disks without the loss of data or disruption in service. This is achieved with the implementation of a spare pool of disks. If one (or several) member drives of an array fails, the global spare drive joins the logical drive and automatically starts to rebuild. Our disk arrays are both read and write cached through flash memory to increase speed. Access speed is homogeneous throughout the file system.

The file system is NFS at the front end and based on ZFS at the back. A high degree of redundancy is built into the system, both on the level of the head nodes (which are dual active/active), and on the level of connectivity (10 Gig Ethernet). Part of the older SAM-QFS based storage systems serve as a backup management system that connects our disk arrays to our tape library. This allows the regular backup of files.

Temporary Scratch Space: /scratch

Note! Scratch space is not backed up. Any files that are kept in the /global/scratch/hpcXXXX area cannot be retrieved if they are deleted.

Note: Scratch is regularly purged. Any untouched data older than 60 days will be deleted.

Scratch space is supplied in the /global/scratch area of the file system. This space is intended for transitory data that are generated during a calculation and are usually deleted shortly after the calculation has finished. However, it is worthwhile to consider keeping other intermediate results that are only needed for a short time on scratch space if there is a danger of exceeding disk quota in /global/home/hpcXXXX or /global/project/hpcgXXXX where hpcgXXXX denotes your CAC groupname.

/scratch is subject to a quota of 5 TB per user. If you require more, please contact us.

Note that currently our scratch space is global, i.e. accessible from all nodes. While this implies slower access than local scratch, it allows data to be used from different nodes within a program run, and it simplifies maintenance.

A directory /global/scratch/hpcXXXX is automatically created when you receive an account. By default, it is only accessible by the owner.

To use the scratch, you will often have to set an application specific environment variable, for instance for the chemistry code "Gaussian":

$ export GAUSS_SCRDIR=/global/scratch/hpcXXXX

Disk quotas

Disk quotas are active on the /home, and /scratch filesets. They are enforced automatically. Once a user exceeds a quota, no further data can be written to the file system by that user, making it impossible to log in in some cases. If this happens, you need to contact us and arrange for freeing up disk space.

You can check your disk usage by using the 'myquota' command.

Note: Some of our users currently exceed our disk quotas on one or several of these areas. For groups that require such larger quotas we can temporarily raise them to allow continuing work. We contact these users regularly to check if such extensions are still required. We do not provide long-term data storage as a default.

If you have special needs concerning disk usage that exceeds the above quotas, please contact us.

Files in /home are automatically backed up. Users do not have to do anything for these activities to occur.

Backups

By default, we maintain backups for the purpose of securing user data (disaster recovery) only, not for permanent storage or external use. This means that it is the responsibility of the individual user to remove data that are to be kept permanently from the cluster and store them on external media, such as disks, tapes, or DVD's.

User data that reside in the /home file system are backed up locally and off-site. This happens automatically.

The backup cycle is the same as for /home. No user data residing outside these directories are backed up. This holds specifically for /tmp and /scratch.

How to get lost data back

contact us. The system administrators will retrieve data from the regular backup. Keep in mind that changes that you made to the data before the loss occurred might be lost since the copy of your file may be outdated. Likewise, if you made accidental changes to your files, you might be able to revert to an earlier version by retrieval from a backup copy. However, if the changes were already committed the earlier file could be lost. To avoid such problems, consider a version control system.

If the loss is the consequence of a general disk failure, the part of the file system that was affected will be restored from safety backups, and it is not necessary (nor useful) to contact the administrator for the retrieval of individual files. In that case, you will have to wait until the file system is restored to normal. This may take several days in the case of a severe failure.

What to do if a lot of disk space is needed

If you need more disk space than the disk quota on /home and /u1 allows, you should consider the following options, preferably in that order:

• Clean up the /home. Delete data that are no longer needed.
• Download permanent data and back them up on external media, such as external drives and arrays, Tapes, or DVD. This is the safest way to make sure your data cannot be lost. We do not supply individual backups or archiving facilities by default. This is the responsibility of the user.
• If the data are temporary (eg, they can then be discarded eventually), they might be written out on /scratch. Data on scratch space are not backed up, and should be removed as soon as they are not needed any more.
• Finally, if you need to keep large amounts of data that are accessed rather frequently, exceed the disk quota, and need to be backed up, you need to contact us and make special arrangements.