It runs into this issue far more consistently. These versions failed less often than R2019a has, but I suspect that's because MATLAB has been getting bigger and bigger with each release. We've also seen similar behavior with versions R2018a and R2018b of MATLAB.
Occurs when the install is run as a Required assignment. It's some strange issue with SCCM that only So I'm 99.9999999% sure it's not an issue with the MATLAB install. I have even set it up to run as a scheduled task as the SYSTEM user using the downloaded files in the cache folder, and it worked fine.
As mentioned before, if you start the install from Software Center as part of a Self-Service assignment, so far it has worked 100% of If you run the MATLAB silent install on your own, it works fine. If you run the MATLAB installer manually, it works fine. The MATLAB setup then simply exits (with no error) and you have a partial, unusable MATLAB install on the machine. However, the uninstaller files don't exist because Java exited before it copied those files. It is supposed to register the uninstaller for MATLAB. The main MATLAB setup either doesn't notice or care that it failed and continues on. Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)Ĭ2:4592860 5627 4 .lucene41.Lucene41PostingsWriter::startDoc (242 bytes)įrom what I can tell, what happens is that the Java installer runs out of memory while copying in the MATLAB program files and exits. T H R E A D -Ĭurrent thread (0x00000000148bb800): JavaThread "C2 CompilerThread0" daemon
Minidumps are not enabled by default on client versions of Windows # Java VM: Java HotSpot(TM) 64-Bit Server VM (25.181-b13 mixed mode windows-amd64 compressed oops) # This output file may be truncated or incomplete.
# Set larger code cache with -XX:ReservedCodeCacheSize= # Decrease Java thread stack sizes (-Xss)
# In 32 bit mode, the process size limit was hit # The system is out of physical RAM or swap space # Native memory allocation (malloc) failed to allocate 629696 bytes for Chunk::new # There is insufficient memory for the Java Runtime Environment to continue. A partial dump of one of those log files is below: When the install fails during a required assignment, a JRE error log file is generated which says that it ran out of memory. MATLAB uses Java during the installation. This Reddit thread shows some other folks running into the same issue, and also the OP mentions that in their experience it only affects Windows 10. I'm wondering if there some difference about how an install is run as Required vs run as Self-Service? Does it have less access to system resources somehow, possibly specifically memory or the paging file?
This problem may also be specific to Windows 10. However, if it's assigned as Self-Service and you start it from SoftwareĬenter, so far it has always succeeded. When the install runs on a machine as a Mandatory/Required assignment, it nearly always fails. The application is Mathworks MATLAB R2019a. Having a problem with an Application Deployment.