Dear all,
I'm currently testing our LSF port of I-Tasser 4.2. However lots of jobs end being stuck in `sync` for hours or even days. I noticed, that many of the scripts end by calling `sync`, sometimes even twice.
What is the rationale behind calling `sync` instead of trusting the underlying OS to "do it right"? `sync` can take a very long time on busy nodes. Our nodes /are/ constantly busy and have lots of RAM.
Did anyone of you ever experience similar problems and how did you solve or work around them? Is there a way to globally disable the `sync` calls of I-Tasser?
Thanks in advance
Frank