We have recently installed the latest upgrade to v4.1.2 only to find that the webapp has been very very slow, such that we have to regularly restart it. This does not appear to be due to users flooding the master node.
We have four nodes with 32 cores, 384 GB memory and 4 GPUs. Quadro RTX 6000.
In addition, the ‘Current Jobs’ section in the ‘Manage Jobs’ tab is empty even when jobs are running, meaning that it is hard to find out who is running what on which GPU node (we have four in total, including the master node).
May I ask whether anyone else is experiencing this issue, and whether there is a patch to fix the Manage Jobs tab, see screenshot below?
Thanks for the post - sorry to hear you’re having trouble using the interface. This is definitely unusual and something that isn’t expected; the v4 interface should be much more responsive than v3.
It’s likely an error occurred during the upgrade process or there’s a network issue causing certain data from being loaded by the browser. To help us debug further, please generate a log of your CryoSPARC instance:
Additionally, you can try editing your cryosparc_master/config.sh file with the following environment variable that modifies what external data the web application interface loads (doesn’t interfere with running jobs):
Stop the instance: cryosparcm stop
Add export CRYOSPARC_DISABLE_EXTERNAL_REQUESTS="true" to the end of your config.sh file