Difference between revisions of "Rlogin"
| (One intermediate revision by the same user not shown) | |||
| Line 4: | Line 4: | ||
Many professors use rlogin as a persistent shell server to run autograder scripts that hit the Canvas API: notably in CS2505, 2506, and Systems. | Many professors use rlogin as a persistent shell server to run autograder scripts that hit the Canvas API: notably in CS2505, 2506, and Systems. | ||
| − | Student use is generally permissive: [[User:Rsk]] used it for bruteforcing hashes before he had access to [[ | + | Student use is generally permissive: [[User:Rsk]] used it for bruteforcing hashes before he had access to [[Infrastructure|LUUG hardware]]. |
== Setup == | == Setup == | ||
| Line 13: | Line 13: | ||
rsk@acidburn:~$ | rsk@acidburn:~$ | ||
</pre> | </pre> | ||
| − | indicates that the rlogin cluster has 37 nodes, 5 of which are GPU boxes. homedirs are mounted from a ceph cluster, as seen below. | + | indicates that the rlogin cluster has 37 nodes load balanced using ipvs, 5 of which are GPU boxes. homedirs are mounted from a ceph cluster and shared between nodes, as seen below. |
<pre> | <pre> | ||
[rishik@hackberry ~]$ df -h | grep ceph | [rishik@hackberry ~]$ df -h | grep ceph | ||
Latest revision as of 23:06, 7 February 2025
rlogin is the Computer Science department's student login cluster. Information about access can be found on its homepage, but at time of writing it is available to declared Computer Science majors, as well as anyone currently enrolled in a CS class.
Use
Many professors use rlogin as a persistent shell server to run autograder scripts that hit the Canvas API: notably in CS2505, 2506, and Systems.
Student use is generally permissive: User:Rsk used it for bruteforcing hashes before he had access to LUUG hardware.
Setup
rsk@acidburn:~$ curl http://rlogin.cs.vt.edu/nodes.txt 2> /dev/null | grep rlogin | uniq | wc -l 37 rsk@acidburn:~$
indicates that the rlogin cluster has 37 nodes load balanced using ipvs, 5 of which are GPU boxes. homedirs are mounted from a ceph cluster and shared between nodes, as seen below.
[rishik@hackberry ~]$ df -h | grep ceph homes@00000000-0000-0000-0000-000000000000.cephfs=/home 77T 20T 58T 26% /home glogin@00000000-0000-0000-0000-000000000000.cephfs=/glogin 77T 20T 58T 26% /opt web@00000000-0000-0000-0000-000000000000.cephfs=/web/export 48T 1.1T 47T 3% /web homes@00000000-0000-0000-0000-000000000000.cephfs=/home/ugrads 20T 11T 10T 51% /home/ugrads [rishik@hackberry ~]$
At time of writing, individual nodes are of the following specs:
- Dell R640 1U server
- 2x Intel Xeon Gold 5218 2.3Ghz 16 core processors
- 384 GB Main memory @ 2933 MT/s
- 1TB system drive
- 10 Gbit Ethernet interconnect
The Rlogin cluster is colocated in McBryde hall.
