I have downloaded the latest version (2.2.1.3) of Cuffdiff/Cuffnorm/Cuffquant on a local instance of Galaxy but keep getting a fatal error: Exit Code 139 and a linked segmentation fault: 11 almost immediately after running these programmes. This occurs even when running a single BAM file. I can see that there has been various discussions on this issue in the past and wondered if this issue has been resolved?
Many thanks
Mark
Hello,
With a new local (yes?) a few issue could contribute this this type of failure. Dependencies missing from the tool install could be part of this as could resources.
Could you share more details?
1. Is this a new local or one that has been configured for production (specifically, the database upgrade to Postgres?) that has executed prior tools successfully with datasets of this size or are these the first jobs?
2. Tools were installed via the Tool Shed with managed dependencies or through other means? When you check through the Admin UI, all tools have a status that reflects functional usage?
3. The version of Galaxy is the current master branch or is the dev branch in use? (v 15.10)
4. What is the size of the data and what type/score of memory (RAM)/CPU resources are dedicated to Galaxy?
5. Other modifications related to production instance upgrades would be helpful to learn about.
There are solution aids we can suggest to help in each case, with some overlap, but this should get us started with troubleshooting. These include alternate community bundled Galaxy distributions with additional pre-configured advanced config (worth using in cases like this).
http://usegalaxy.org/toolshed
http://getgalaxy.org -> start with the default set up on that wiki and continue to production config. Not all are upgrades are necessary for personal use locals, but some key portions (such as the database upgrade) can make a difference with compute intensive tools.
Thanks, Jen, Galaxy team