All services start up fine - gsad, gvmd, redis, and openvas-scanner
I can log into the UI - I can run any of my discovery scans and they will run to completion - but when I attempt to run a vulnerability scan it quits at 1%. The following is logged in the gvmd.log
event task:MESSAGE:2021-09-01 15h41.30 UTC:25816: Task ap01mgt103 (56a2b982-4a16-4c24-9b7e-9f0fb08c2506) has been requested to start by bbarendt
event task:MESSAGE:2021-09-01 15h42.08 UTC:25826: Status of task ap01mgt103 (56a2b982-4a16-4c24-9b7e-9f0fb08c2506) has changed to Running
md main:WARNING:2021-09-01 15h42.13 UTC:25826: openvas_scanner_read: Failed to read from scanner: Connection reset by peer
event task:MESSAGE:2021-09-01 15h42.14 UTC:25826: Status of task ap01mgt103 (56a2b982-4a16-4c24-9b7e-9f0fb08c2506) has changed to Interrupted
and the following is logged in messages
openvassd[1342]: openvassd: /builddir/build/BUILD/openvas-scanner-6.0.0/src/pluginscheduler.c:87: plugin_add: Assertion `category >= ACT_INIT && category <= ACT_END’ failed.
Those are quite outdated components of the GVM stack which for a GVM branch which is also even EOL and unsupported. AFAIK the Atomic Repo is only providing current / up2date version of GVM for CentOS 8 (See GitHub - Atomicorp/gvm: Greenbone Vulnerability Manager / Openvas packaging project) so an upgrade to CentOS 8 and an upgrade of the GVM stack afterwards is recommended which should solve this issue.
I am quite aware that they are old - but this is a community forum - so I was looking for a community forum answer from others who have had the same issue. Since one does not seem to be forthcoming - I have opted to abandon GVM and go with Tenable. Perhaps when this program matures into a reliable product, I will revisit it.
Looks like a misunderstanding is happening here. For completeness:
The used packages for CentOS 7 are originating from a 3rdparty repository which has already dropped support for CentOS 7.
Furthermore the packages are lacking multiple patch releases of the GVM components (e.g. gvmd is at version 8.0.0 while version 8.0.2 was released more then a year ago).
Based on the usage of such outdated version (which already have received multiple patch releases) the only valid answer (also from community side) could have been:
Please update to more recent versions as the chances are very high that this problem has been already fixed.
I have 6 GVM systems running - 2 are still working, 4 are not. For all intents and purposes everything is working in that all the processes are running without any errors - the only error occurs when trying to run a vulnerability scan - a discovery scan works just fine. I am assuming it is because the discovery scan does not use any plugins. The systems were running a scheduled monthly vulnerability scan just fine, then in June the scans failed with no apparent cause. The only common thread for the 4 that broke is they are VMs running CentOS 7 in Azure - the 2 that still work, are VMs running CentOS 7 on VMWare, yet they are all running the exact same versions. So I don’t think the version of what I am running has ANYTHING to do with why it is NOT working.
I can remember that the openvas-scanner had various problems / bugs in it’s internal plugin scheduler triggered by some race conditions (e.g. depending on the exposed service of a target which is influencing the launch order of VTs).
And these problems / bugs had caused such problems (the initially posted error message mentions pluginscheduler.c) but already got fixed in more recent released versions of the scanner.