The “node” command running under the cryosparc user account normally fluctuates between about .3 and 12% CPU as seen by top. Sometimes our cryosparc web interface seems to hang and when I check, the "node’ process is at 100%. Only solution is to restart cryosparc. “node” command returns to normal state. This is a new issue with this Crysoaprc 3.3.2 server, been running for over a year. Thanks, Doug
We hypothesize that spikes in the node process’s CPU usage could be caused by a network disruption. Does $CRYOSPARC_MASTER_HOSTNAME point to non-loopback address in your case?
To find out, run in a fresh shell
eval $(/path/to/cryosparc_master/bin/cryosparcm env)
host $CRYOSPARC_MASTER_HOSTNAME # check output of this command
exit # avoid running additional commands in this shell
Where there recent network disruptions, usage spikes or a reconfiguration of the network?
1 Like
Yes, non-loopback. Network traffic can be quite high at time, but this is the only one of our 11 cryosparc servers exhibiting this issue, but it is the most heavily used. From what I have observed the “node” process goes to 100% and stays, a restart of cryosparc brings back down to normal range, in-line with our other cryosparc servers. I will look for net work related issued next time we observe a spike in the “node” process. Thanks,