On 4 December 2010 15:49, Dev Null <[email protected]> wrote: > Is there any specific debugging tool or techniques available to debug highly > critical production server, where data might be corrupted or OS is behaving > crazy?
If you have source code access to the kernel and know the functions that you would like to monitor, you can use kprobes. You'd have to write a kernel module and attach to the function that you wish to debug. Since it's a kernel module, you'd have to be careful about semantics. http://lwn.net/Articles/132196/ tcpprobe is a wonderful example that uses kprobes. http://www.linuxfoundation.org/collaborate/workgroups/networking/tcpprobe Its source code is here: /path/to/linux-$(uname -r)/net/ipv4/tcp_probe.c Good luck, -- Vimal -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to [email protected] Please read the FAQ at http://kernelnewbies.org/FAQ
