Because datasets share all available space in a pool, we use multiple datasets to maximize our flexibility in storing Colleague data. Each Colleague environment has its own dataset, and the daemon, Colleague, and LPR databases (daemon, coll18, and repository das) each have their own datasets as well.
In addition to Colleague itself, we also have the Unidata binaries stored on ZFS, so both /usr/ud73 and /usr/unishared are mounted as ZFS datasets. We also have an extra dataset, vol1, that we can use to store backup data such as crontabs and config files usually stored in /etc.
The resulting ZFS list looks like this:
| NAME | USED | AVAIL | REFER | MOUNTPOINT |
| m10-1 | 345G | 749G | 31K | /m10-1 |
| m10-1/coll18 | 47.0G | 749G | 46.6G | /datatel/coll18/coll18 |
| m10-1/daemon | 6.03M | 749G | 3.71M | /datatel/coll18/daemon |
| m10-1/development | 2.03G | 749G | 88.0G | /datatel/coll18/development |
| m10-1/production | 284G | 749G | 92.2G | /datatel/coll18/production |
| m10-1/repository das | 51.7M | 749G | 48.6M | /datatel/coll18/repository_das |
| m10-1/test | 11.3G | 749G | 91.8G | /datatel/coll18/test |
| m10-1/ud73 | 650M | 749G | 640M | /usr/ud73 |
| m10-1/unishared | 2.25M | 749G | 1.90M | /usr/unishared |
| m10-1/vol1 | 140M | 749G | 139M | /vol1 |
You may have noticed that while all three environments refer to the same amount of data (~90GB), the development and test environments actually use only a fraction of that. That's because both were cloned from production, and the dataset only uses space to write changes since the clone was made (January for test and last September for dev). The Colleague clone process is one where we can truly leverage ZFS features.