INLA under R complains about GLIBC.

Running INLA under R can give the error "/lib64/ version `GLIBC_2.23' not found" or similar. A workaround is to issue INLA:::inla.dynload.workaround() before running code using INLA.

INLA as it is distributed has certain expectations of the base system that RHEL7 currently does not meet. The error will typically look something like this:

res.inla <- inla(y~1+f(time,model="ar1",
+                           hyper=list(theta1=list(param=c(mtauprior$shape,1),initial=mtauprior$initial),
+                                    theta2=list(param=c(rhoprior$mu,1),initial=rhoprior$initial))),
+                            control.fixed = list(mean.intercept=alphaprior$mu,prec.intercept = alphaprior$prec),
+                            family=fam,data=d.inla,
/mn/sarpanitu/modules/packages/R/3.5.0-gcc/lib64/R/library/INLA/bin/linux/64bit/inla: /lib64/ version `GLIBC_2.23' not found (required by /mn/sarpanitu/modules/packages/R/3.5.0-gcc/lib64/R/library/INLA/bin/linux/64bit/inla)
/mn/sarpanitu/modules/packages/R/3.5.0-gcc/lib64/R/library/INLA/bin/linux/64bit/inla: /lib64/ version `GLIBC_2.23' not found (required by /mn/sarpanitu/modules/packages/R/3.5.0-gcc/lib64/R/library/INLA/bin/linux/64bit/
Error in inla.inlaprogram.has.crashed() : 
  The inla-program exited with an error. Unless you interupted it yourself, please rerun with verbose=TRUE and check the output carefully.
  If this does not help, please contact the developers at <>.

Unfortunately, upgrading GLIBC is not an option, but drift@math will see if we can rebuild INLA to resolve this issue. For now though, issuing INLA:::inla.dynload.workaround() is the best option.

Published July 26, 2018 3:05 PM - Last modified July 26, 2018 3:05 PM