bcc-tools: trace.py syscall trace not working for proccess in container(lxc)
- Dominant language
- C
- Stars
- 22.7k
- Forks
- 4.1k
- Avg merge
- 10d 4h
- Merged PRs (30d)
- 3
Description
im trying to debug a init proccess in a lxc container. for better reproduce, im using `mknod` tool instead bellow
with `trace 'c:mknodat "%s", arg2'` i can see the call to `mknodat` while im running `mknod test c 1 11` out of container.
but if i run the same thing in container. trace.py didnt show me anything. by using strace, i confirmed the `mknodat` syscall has been send out without any issue.
is this a expected behavior?
im using `trace 'filename_create(int dfd, struct filename *name, struct path *path, unsigned int lookup_flags) "%s", name->name' -UK` (and find through stack manually) as a alternative now. because `do_mknodat` not exported and optimized by compiler, which changed signature
version: 5f21a7274
kernel version: 6.8.1
Contributor guide
No contributing guide indexed for this repository
Research direction
Start with trace.py and reproduce the mknodat case on kernel 6.8.1 both inside and outside the LXC container, comparing BCC output with strace. Determine whether the missing trace is expected and identify the relevant attachment or filtering behavior; done means the behavior is explained and a concrete correction or documentation outcome is defined.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- linux, python
- Domain
- observability, operating-systems, tooling
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 30/100