Message boards :
ATLAS application :
Changes to use configured proxy for Frontier servers
Message board moderation
| Author | Message |
|---|---|
|
Send message Joined: 13 May 14 Posts: 387 Credit: 15,314,184 RAC: 0   |
With big thanks to computezrmle (who provided the code), ATLAS WU will now use your HTTP proxy (if configured in BOINC) to access the Frontier servers from which database information is read during the task. This means that you should see traffic going through your configured proxy instead of http://db-atlas-squid.ndgf.org:3128. The new WU with this change are just going in now, so it may take some time before it has an effect. It works for both vbox and native tasks. As always, let us know of any issues you see. |
 Send message Joined: 15 Jun 08 Posts: 2606 Credit: 262,449,156 RAC: 136,877 |
ATLAS native works as expected. ATLAS vbox needs to reconfigure the VM's CVMFS. It's still a todo to modify the startup scripts accordingly. |
|
Send message Joined: 9 Dec 14 Posts: 202 Credit: 2,533,875 RAC: 0 
|
for native atlas the squid logs show connections to http://lcgft-atlas.gridpp.rl.ac.uk:3128 which is one of the frontier ATLAS servers. Seems like using the proxy is a good idea since it shows 100% hitrate for this site (not sure if that makes sense). |
|
Send message Joined: 15 Jun 08 Posts: 2606 Credit: 262,449,156 RAC: 136,877 |
Indeed, lcgft-atlas.gridpp.rl.ac.uk:3128 is also a squid proxy (better: an array of squids). Here it is configured to hide/speed up the original backend systems. 100% hitrate sounds quite normal. Although this will drop a bit over the day as all files are marked with a max age date. Hence they will be refreshed after that date has expired. |
©2025 CERN