I would like to thank the people who offered suggestions in response to my query about installing the latest Topspin under CentOS 7. Eventually, I found everything worked fine if I installed Topspin in the default /opt path, probably without any more preparation needed than installing the rpm's contained in the Topspin operating and acquisition environments as described in the instructions, and if I used any other path, it wouldn't work. I have a suspicion about why that is, but having gotten the system working after spending much more time with it than I planned, I am not starting over to test the idea.
I have been updating our Varian systems as well to OpenVnmrJ running under CentOS 7. A generic installation of CentOS will turn on SELinux by default, but in the OpenVnmrJ installation script, it gets turned off. When I got the Topspin installation working in the /opt tree, I had immediate problems with using Samba for read-only network drives for users to retrieve their data from the /home tree--key files were being hidden on the network drives even though their non-SELinux permissions were set for them to be visible. This problem didn't exist in the Varian installations, and it went away when I disabled SELinux in the Bruker installation. It's just a guess that SELinux settings caused my original problem.
A few people asked if I could be having a problem with the CodeMeter licensing, which is being used for Topspin >3.6. This wasn't the case, and in my experience with it, CodeMeter is more flexible than Flexlm and a significant improvement on it. Trading in my Flexlm license was straightforward, and if you have been thinking of upgrading your Topspin but unsure of how much hassle the new licensing might be, hey, it's the lowest speed bump you will need to clear. I've been horsing around with Flexlm for 20 years I think; farewell, Flexlm!
Along with CodeMeter, I would like to put in a plug for OpenVnmrJ. I felt it was extremely beneficial to have paid to join the short-lived OpenVnmrJ software consortium run by Dan Iverson and Krish Krishnamurthy. We will be running Varian instruments here well into the future, and this was an opportunity to get the port to CentOS 7 done and a few things fixed in VnmrJ (or at least a convincing explanation of why they couldn't change, if they are in the proprietary console software). I highly, highly recommend the OpenVnmrJ running under CentOS 7--in key respects, I have never been happier with the Varian software than I am right now. On our busy automated spectrometers with many operators logging in and out all day, we had some stability problems with VnmrJ 3, that got much better with VnmrJ 4, and OpenVnmrJ seems to be super stable. If you are already running VnmrJ 4, there's basically no learning curve to adapt to OpenVnmrJ. If I were running an earlier VnmrJ, I would want to be getting as close to the last version as I could, while the software is still being freely distributed and so is the operating system it depends on.
The consortium also included a chance to get Krish's CRAFT box which he has extended to analyze Bruker data as well (including, as of recently, Bruker NEO data). I attended the post-SMASH CRAFT workshop last September, as did a lot of SMASH attendees, and I am really sold on this approach to quantitation.
David VanderVelde
Manager, Liquids NMR Facility, Caltech
davidv_at_caltech.edu
Received on Wed Jan 23 2019 - 04:34:14 MST