What were you trying to do that didn't work?

Collect coredump files using SoS plugin that are larger than 200M

What is the impact of this issue to you?

Customer requires to collect coredumps of their application using the SoS feature

Please provide the package NVR for which the bug is seen:

since RHEL-9.2.0 when the coredump.py plugin was introduced
package: sos-4.10.0 and later revisions in RHEL 9

How reproducible is this bug?:

Always with coredumps larger than 200M

Steps to reproduce

1. Create a set  of different coredump sizes 1G, 200M, 400k
2. dd if=/dev/urandom of=/tmp/random1g bs=1024 count=1048576
3. dd if=/dev/urandom of=/tmp/random200m bs=1024 count=208000
4. python -c 'with open("/tmp/random1g", "br") as fp: foo = fp.read(); print("1GiB Read Into Memory"); import os; os.abort()'
5. python -c 'with open("/tmp/random200m", "br") as fp: foo = fp.read(); print("200M Read Into Memory"); import os; os.abort()'
6. python -c 'import os; os.abort()'
7. Run the sos tool with below options:
8. sos report -vv --all-logs --log-size=0 --only-plugins=coredump --batch

Expected results

Collection of all coredump succeeds

Actual results

Collection of the coredump that is larger than 200M gets skipped

[plugin:coredump] Skipping core dump file /var/lib/systemd/coredump/core.python.0.546e2ce70a254227bdca9f19590228ac.6789.1764277047000000.zst due to size

Additional Information

The customer tracked the issue to the coredump.py plugin itself, it shows a hardcoded limit 

 86                     # a_c_s does not return any information for a skipped file,
 87                     # so stat the size here and if the core is larger than our
 88                     # limit, move on to the next
 89                     # TODO: do not hardcode this. Extend log-size to per-plugin
 90                     # TODO: option and link this to that value
 91                     if os.stat(core_path).st_size > 209715200:
 92                         self._log_info(
 93                             f"Skipping core dump file {core_path} due to size"
 94                         )
 95                         continue

In upstream code this may be a regression from being refactored, commit: 99d0ae2
For our code the plugin was included with this issue.