Monitoring disk usage

First check the existing usage

To check the amount of disk space available, e.g. on /data/server_name

cd /data/server_name         e.g. cd /data/blackbird

df -h .

will produce the following output:

Filesystem

Size

Used

Avail

Use%

Mounted on

/dev/mapper/vg_blackbird-lvdata

2.6T

1.5T

1017G

59%

/data/blackbird


which shows that /data/blackbird is currently 59% full and 1017 GB is available.

Next, make sure we won't break the 90% full rule

In order for the disk to function effectively we must make sure that it is never more than 90% full.
So if you have not used any space there already, the space available is the smaller of:

applying the one third rule: 2.6 TB / 3   ~= 860 GB

applying the < 90%-availability-use-rule: 1017G-(2.6 TB / 10) ~= 770 G

so if you needed more than 770 GB of space, please wait or try another server.

Don't forget your existing usage

However if you have already used space, you need to subtract your current usage, e.g.

cd /data/server_name/your_user_name
du -h -s .
27G.

and

cd /data/server_name/not-backed-up/your_user_name
du -h -s .
273G

Adding these two together gives us 300 GB, which needs to be subtracted from the smaller of the two figures above, so 770 - 300, leaving 470 GB of space for your jobs.  You should of course move any data no longer being actively used once your jobs have finished.

However as multiple users are likely to be using the system, somebody else may also consume resources whilst your jobs are running, which is why we ask you to monitor usage closely and if necessary kill jobs before the system runs out.