-
Notifications
You must be signed in to change notification settings - Fork 3.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
multiple concurrent probe users, fixing "Device or resource busy" #872
Comments
@brendangregg
Here is my diff file to fix this.
|
G'Day @derek0883 , thanks for this, this is important to get done. I'm traveling at the moment so haven't tested it yet. @drzaeus77 / @4ast : did you have an opinion about solving this in python vs C? |
adopting instances is certainly a good move. I'm not sure adding _pid is enough. May be add _bcc suffix or something as well. I think the major bits should be on C side. If we need to change the C api, let's do it. |
@4ast |
_pid is probably enough from uniqueness point of view. I was thinking to improve debugging... so doing 'ls' in instances will be clear what tool created them. |
@brendangregg @4ast
|
great! yes. please send PR. It's easier to review via PR :) |
fixed by #918 |
bcc currently has a limitation of one program per probe. Eg, you can run execsnoop & opensnoop at the same time, but not two opensnoops, as the second gets:
As bcc/BPF gets adopted, I'm seeing the risk of clashes between 24x7 bcc/BPF monitoring and ad hoc analysis, when the latter tries to instrument the same probes.
I suspect it's fixable in just bcc, without kernel changes. For ftrace, anyway, we can create two probes for the same kprobe location, provided they have different aliases, and trace_pipe will see both events:
The text was updated successfully, but these errors were encountered: