Difference between revisions of "Getting Started - File Structure"
From Xem2
Jump to navigationJump to search (Created page with "==File Structure at JLab== ===Overview=== Users can access the JLab Common Users Environment (CUE) by tunneling through login1 or login2 server at JLab. <code> ssh -t -L6437:...") |
|||
| Line 13: | Line 13: | ||
====Using Putty to SSH==== | ====Using Putty to SSH==== | ||
| − | ====Common Connection Issues | + | ===File Structure=== |
| + | * /group/ | ||
| + | * /work/ | ||
| + | * /cache/ (jcache utility and scicomp docs) | ||
| + | * /volatile/ | ||
| + | * farm_out | ||
| + | * home dir | ||
| + | |||
| + | ===Using JLab Farm=== | ||
| + | *Auger | ||
| + | **Practice submitting stuff... | ||
| + | *SWIF | ||
| + | **Practice submitting stuff... | ||
| + | *hcswif | ||
| + | '''Common commands and difficulties with jobs''' | ||
| + | |||
| + | ===Common Connection Issues=== | ||
* Tunneling Problems | * Tunneling Problems | ||
** Conflicting or incorrect SSH keys in known_hosts | ** Conflicting or incorrect SSH keys in known_hosts | ||
| Line 21: | Line 37: | ||
** VNC Servers started on different destination_computer are not accessible if you login later to some other destination_computer. | ** VNC Servers started on different destination_computer are not accessible if you login later to some other destination_computer. | ||
** The computer was restarted, so the vncserver was shut down. | ** The computer was restarted, so the vncserver was shut down. | ||
| − | * JLab /group/ /work/ /volatile/ Directories Missing | + | * JLab /group/ /work/ /volatile/ Directories Missing (Hanging) |
** Sometimes work on connection gateways will break Lustre file system connections that are not tier-1 | ** Sometimes work on connection gateways will break Lustre file system connections that are not tier-1 | ||
** Improper JLab farm node use can cause the Lustre file system to hang. Report an incident report using the [//https://jlab.servicenowservices.com/nav_to.do?uri=%2Fhome.do%3F Incident Report Utility] | ** Improper JLab farm node use can cause the Lustre file system to hang. Report an incident report using the [//https://jlab.servicenowservices.com/nav_to.do?uri=%2Fhome.do%3F Incident Report Utility] | ||
Revision as of 12:51, 30 August 2021
File Structure at JLab
Overview
Users can access the JLab Common Users Environment (CUE) by tunneling through login1 or login2 server at JLab.
ssh -t -L6437:localhost:5912 cmorean@login2.jlab.org ssh -t -L5912:localhost:5912 <destination_computer>
- The red numbers should NOT be 12 for you. (That may cause our vncservers to conflict if we tunnel to the same destination_computer).
- It is common to connect to JLabl1 - JLabl3, or an ifarm computer.
- In order to use a vncserver Highly recommended over x11 port forwarding, you will need to use a computer other than an ifarm computer.
You can check your connection speed to JLab is not the problem by using JLab's SpeedTest.
VNC Server Setup
Using Putty to SSH
File Structure
- /group/
- /work/
- /cache/ (jcache utility and scicomp docs)
- /volatile/
- farm_out
- home dir
Using JLab Farm
- Auger
- Practice submitting stuff...
- SWIF
- Practice submitting stuff...
- hcswif
Common commands and difficulties with jobs
Common Connection Issues
- Tunneling Problems
- Conflicting or incorrect SSH keys in known_hosts
- VNC Server Problems
- If you tunnel and only specifying login.jlab.org, the machine will choose either one or the other, which sometimes breaks vncservers for UNKNOWN REASONS.
- VNC Servers should be started on the port highlighted in red using the command <vncserver -geometry 1920x1080 :<port_number>
- VNC Servers started on different destination_computer are not accessible if you login later to some other destination_computer.
- The computer was restarted, so the vncserver was shut down.
- JLab /group/ /work/ /volatile/ Directories Missing (Hanging)
- Sometimes work on connection gateways will break Lustre file system connections that are not tier-1
- Improper JLab farm node use can cause the Lustre file system to hang. Report an incident report using the Incident Report Utility
