scancode job will be killed about fossology HOT 9 OPEN

This is easy to fix. Change this line in scancode.conf:
special[] = NOKILL

from fossology.

After upgrading to 4.4.0 (installation from sources)I encountered the same issue, but randomly: it happens roughly on 1 out of 10 uploads or so (both big and small, so upload size doesn't seem to matter). I applied the suggested fix, and now it seems to work. I guess that this issue should be investigated a little bit more. What could I do to provide more debug info?

from fossology.

Use my patch above and it will not be killed any more.
I am using now an automated script that patches the docker image after installation (with multiple fixes).

from fossology.

yes @uholeschak but IIUC this means that the scancode agent does not do something that it is supposed to do, or that something else wrong is happening so that the scheduler believes the agent is hanging and needs to be killed. So your workaround works but IMHO it would be better to find and fix the issue

from fossology.

No this is not a bug, simply your system is to slow and a timeout occurs.

from fossology.

16 CPUs 3.3MHz, 32GB RAM, fast SSD... I keep monitoring everything with top, the CPU used by scancode is always at 100%, but that happens also with nomos, which is even slower but it doesn't get killed 🤔

from fossology.

I expect the watchdog processing in scancode is buggy. If the summary time is to long it always gets killed.
It seems that the alive handling in scancode is not working.
The timeout is fixed and could not be changed only switched off.
Another solution would be keeping the uploads small.

from fossology.

Use my patch above and it will not be killed any more. I am using now an automated script that patches the docker image after installation (with multiple fixes).

@uholeschak could you share the script? It would be super-useful! Thanks 🙂

from fossology.

Basically the bat is like this (only the scancode patch in this case):
The chmod fixes an additional rights problem,

SETLOCAL EnableDelayedExpansion

set "BatDir=%~dp0"
set "ContainerId=%1"

IF "!ContainerId!"=="" (
  echo missing container id
  EXIT /b 1
)

docker exec -i "!ContainerId!" sh -c "chmod a+xrw -R /root" || EXIT /b 1

docker cp "!BatDir!\scancode.conf.diff" "!ContainerId!:/usr/local/etc/fossology/mods-enabled/scancode/scancode.conf.diff" || EXIT /b 1
docker exec -i "!ContainerId!" sh -c "cd /usr/local/etc/fossology/mods-enabled/scancode && patch -N < /usr/local/etc/fossology/mods-enabled/scancode/scancode.conf.diff"

scancode.conf.diff:

--- scancode.conf	2023-11-22 10:38:18.000000000 +0100
+++ scancode_mod.conf	2023-12-08 17:52:33.757401300 +0100
@@ -20,4 +20,4 @@
 ; A comma separated list of values.
 ; Directives:
 ;     EXCLUSIVE: the agent cannot run concurrently with any other agent.
-special[] =
+special[] = NOKILL

from fossology.