I have met a prolem that I cache the particles to ssd become slower than others on a cluster.My speed always <100M/s and others always 300M/s.I don’t know how come to this problem.Because maybe we use the same ssd at different time,but we didn’t have the similar cache speed.I think it’s very weird.Looking forward your advice.Thanks very much.
And I also have 300M/s before my cryosparc stop accidently when I memory is full.Now I have restart the cryosparc and the memory isn’t full,but the speed didn’t recover.
I have the same problem too. On good days we have speeds between 250-300 MB/s but now and then speeds drop to less than 20 MB/s.
This happened today on two workers (two independent jobs/projects) reading from their internal raid0 drive to their internal NVMe drive for cache.
Is it a problem on the Master or database? We recently moved our cryosparc database to be on a local SSD on the master rather than on a network resource.
Any ideas to monitor/diagnose the source of the problem would be great.
Hi Arnaud, have you been able to find any correlations between job type/hours/system load or anything and the variation in ssd caching speed? Did you notice the variation is caching rates when your database was stored on the network?