Hi, I checked the log file of my completed job (syn7829124). There are many dcm files can not be found. I am not sure what's the problem. I think they should be there, otherwise, they can't be read into the subject list in the label file. Thanks!

Created by Chao Huang chaohuang
This was my mistake: Your submission started running again on the machine that we earlier learned was missing images. In response to your post we took the machine off line again (this time ensuring it will not be restarted) and restarted your submission (7862222) on another machine.
hello i encountered this error again: /trainingData/146747.dcm not found, this is submission id: 7862222 i found the missing images are very different on every machine, where the program dies in such an unpredictable way, can you please check?
> Will this consuming time be added into our total computing time? No. We compute the quota you used by subtracting the time the submission started from the current time. When we restart your submission the start time is reset so you are not penalized for the time the submission ran previously.
Thanks. But it takes us several hours to run on that server. Will this consuming time be added into our total computing time?
Dear @chaohuang: We determined there was an issue with the server on which your submission (7837056) was running. The server has been taken off-line for analysis and your submission has been restarted on another server. Sorry for the inconvenience.
Hi Chao, As Yuanfang Guan, we are tracking this issue [here](https://www.synapse.org/#!Synapse:syn4224222/discussion/threadId=1347).
that was my question that i have reported it for like three times since last month.. i think about 0.2% of the images are missing. see my other thread https://www.synapse.org/#!Synapse:syn4224222/discussion/threadId=1347 and it is a different set that is missing on every machine. because sometimes, a image can be found, other times, the same cannot be found, every time it is different. so anyway the result can not be deterministic since the training set maybe even teh testing set is completely different.

dcm files can be found when running preprocessing pipeline page is loading…