I am using RHEL 5.10 servers ( which are in cluster using some shared services). When I restart them using "init 6" or "reboot". the servers hang while shutting down, continuously showing the below message:
sock_sendmsg-error 22
I've to get in the ilO cli mode and do a power reset after which the servers again restart properly. I don't see any significant error in /var/log/messages after that. Can someone expalin what the error means. Is it some problem with kernel level sockets. ?
22 will be a standard errno value. These are defined in a series of (many) C-language include headers starting with errno.h. I find the easiest way of looking these up is with a tool such as cscope. In any case, on most Linux machines, 22 is the value corresponding to EINVAL, which has the standard rendering of "invalid argument".
#define EINVAL 22 /* Invalid argument */
"/usr/include/asm-generic/errno-base.h"
Since we know this to be related to the sendmsg system-call, we can look up the manual page for sendmsg(2), which just tells us that an invalid argument was passed to it, perhaps because something didn't handle an error gracefully. This can also bubble up from other error, such as when the mount command is mounting a CIFS network share and there is an error such as invalid credentials. In cases such as this, the true error may be found in the kernel log (dmesg).
I'm not sure what system component would be reporting that error though. Hopefully you'll see it in the logs also?
