Hi, I have a windows server running orthanc 1.6.0, and encountered an issue with extreme memory usage by Orthanc, whilst seemingly being idle.
General specifications:
Orthanc version: 1.6.0
OS: Windows 10 64 bit, version 1909
Storage directory: Dedicated drive local to the server
Database: PostgreSQL running on the local server
No custom functions/scripts, aside from a LUA script to auto-route instances ‘on stable study’
Uptime on the orthanc server to the time of the issue was just over one month.
Issue:
For some reason, the othanc.exe process suddenly and unexpectedly consumed all available RAM and the entire paging file, crashing windows.
The server only has a small amount of physical ram (8GB), but orthanc generally seems to consume a minute amount of memory.
There were no queries or new instances added to the server within a several hour time-period prior to this event.
at 04:55 on 17/07, windows event viewer logged the following from ‘Resource-Exhaustion-Detector’;
`
Windows successfully diagnosed a low virtual memory condition. The following programs consumed the most virtual memory: Orthanc.exe (752) consumed 30333353984 bytes
`
This equates to just over 30GB of ram consumed. The orthanc logs show nothing prior or during this event, and only these events a couple of minutes afterwards - I’m assuming PostgreSQL crashed due to absolutely no available memory to allocate.
`
E0717 04:57:52.516967 PluginsManager.cpp:164] PostgreSQL error: server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
E0717 04:57:52.529946 PluginsManager.cpp:164] The database is not available, closing the connection
E0717 04:57:52.529946 PluginsManager.cpp:164] PostgreSQL error: no connection to the server
`
Is there any further investigation I can do into the cause of this?
I looked at the changelog for 1.6.1+ and couldn’t see anything to point to this being a known issue.
Cheers
Reuben