I am attempting to run RnBeads on my institutions Cluster, and I keep getting stuck at the same step yielding this error:
Loading required package: doParallel
/usr/bin/gs: error while loading shared libraries: libjasper.so.1: cannot open shared object file: No such file or directory
Error in logger.error(em) :
Could not create file. error in running command
Calls: rnb.run.analysis ... tryCatchList -> tryCatchOne -> <Anonymous> -> logger.error
In addition: There were 16 warnings (use warnings() to see them)
Execution halted
It seems that ghostscript (/usr/bin/gs) cannot locate libjasper.so.1, so I checked the linked libraries for gs:
> ldd /usr/bin/gs | grep libjasper
libjasper.so.1 => /usr/lib64/libjasper.so.1 (0x00000034cf200000)
I found libjasper.so.1 in /usr/lib64/ using:
>ls -la | grep libjasper
lrwxrwxrwx 1 root root 18 May 15 12:00 libjasper.so.1 -> libjasper.so.1.0.0
-rwxr-xr-x 1 root root 339672 Jan 22 2015 libjasper.so.1.0.0
It appears that libjasper.so.1 is symbolically linked to libjasper.so.1.0.0. After doing some research I think that libjasper.so.1.0.0 is a newer debugged version of libjasper.so.1.
My question is: why can gs not find libjasper.so.1? Can it somehow not follow the symbolic link to libjasper.so.1.0.0, or could this have to do with the fact that I am not Root and don't have access to this library? If so, why can I access /usr/bin/gs which is also only accessible by root?
Thank you in advance for any insight you can provide me with.
Seems like this is really a problem with an incomplete installation of ghostscript and nothing to do with RnBeads or Bioconductor. Can you install ghostscript in the approved way for your distribution, for example
sudo apt-get install ghostscripton Ubuntu/Debian?
Also, you might try and figure out what "gs" command line RnBeads is calling and use that as your minimal reproducible test to figure out if ghostscript is installed correctly.
Then you will also be able to ask in the appropriate place (some ghostscript forum? or stackoverflow?) a focused, minimally reproducible question if you still have problems, without needing to confuse the issue by mentioning Bioconductor.
It turns out the problem was in fact due to the Jasper-libs package not being installed on the compute nodes I was using. I am new to using the shared computing cluster, so I was not aware of the possibility that the compute nodes would have different software than the nodes where I was setting up my analysis (it seems the shared software is in fact supposed to be the same on all nodes, but this discrepancy was an error).
I apologize if this is the wrong forum for this type of question (as it appears to be). Sometimes I find it difficult to tell whether the issue is due to Bioconductor software or due to some issue on the cluster. The IT support was taking a very long time to respond so I thought I would give it a shot here.
Thank you for your response all the same.
FWIW, the question seems appropriate for the forum (helping you rule out Bioconductor as the problem), even if in the end the solution involves system administration.