On 07/07/17 11:46, Sorin Gheorghiu wrote:
Hi,after using fuseki for almost one year long, today it failed to start due to insufficient memory:
So what changed? The configuration of the host OS maybe?
# /etc/init.d/fuseki start Starting fuseki...OpenJDK 64-Bit Server VM warning: INFO: os::commit_memory(0x00000005c0000000, 8589934592, 0) failed; error='Cannot allocate memory' (errno=12)# # There is insufficient memory for the Java Runtime Environment to continue.# Native memory allocation (mmap) failed to map 8589934592 bytes for committing reserved memory.# An error report file with more information is saved as: # /opt/apache-jena-fuseki-2.4.0/hs_err_pid21259.log done.
This error is the JVM crashing ... this should never happen and isn't something that is caused by Jena or indeed any java code at all. There is something wrong with the machine. It shouldn't even happen if you give the wrong heap settings - you should get a graceful error message.
As to why heap changes have an effect, I don't know. The error is about mapped file space, not the heap.
Looks like the underlying host OS needs a change to allow the VM the space it wants. You need to talk to the sys admin for the hardware.
Andy
It seems this happens because of small *vm.max_map_count* value 65530 (Elasticsearch has a similar problem). While the server is running on Proxmox as VM server, this value seems to be unchangeable (as root):# sysctl -w vm.max_map_count=262144 sysctl: permission denied on key 'vm.max_map_count' # echo 262144 > /proc/sys/vm/max_map_count -bash: /proc/sys/vm/max_map_count: Permission denied Only decreasing the Java heap size (-Xmx/-Xms) from 12Gb to 8Gb did help. # /etc/init.d/fuseki start Starting fuseki... done.Could you provide an feedback about this error? (the error report file is attached)Thank you, Sorin
