Re: init scripts problems ?

"Also sprach denis bonnenfant:"
> > But the clients are still there, no? 
> 
> no they are all dead :
> 
> Imean :  i kill them with "kill 4049", and they disappear from ps ax output.

OK - then the clients are not blocked.

> ldap:~# tail trace4.4050
> rt_sigprocmask(SIG_UNBLOCK, [ALRM], NULL, 8) = 0
> rt_sigprocmask(SIG_BLOCK, [ALRM], NULL, 8) = 0
> gettimeofday({1134046000, 979532}, NULL) = 0
> rt_sigprocmask(SIG_BLOCK, [ALRM], NULL, 8) = 0
> setitimer(ITIMER_REAL, {it_interval={0, 0}, it_value={60, 0}}, NULL) = 0
> rt_sigaction(SIGALRM, {0x80559c0, [ALRM], SA_RESTART}, {SIG_IGN}, 8) = 0
> rt_sigprocmask(SIG_UNBLOCK, [ALRM], NULL, 8) = 0
> rt_sigprocmask(SIG_UNBLOCK, [ALRM], NULL, 8) = 0
> ioctl(4, 0x4004ab0b, 0xb7e42008) = -1 ETIME (Timer expired)

That's interesting. I wonder what that is. ab0b?


     #define MY_NBD_GET_REQ       _IOW(0xab, 0x0b, char *)

Oh, OK. The clients were waiting to see if anyone needed them to do
anything. Nobody did.

> +++ killed by SIGKILL +++

OK. Not SIGTERM?? Why not?

> but module remains used, and raid locked...

The module IS used - by raid. There is nothing wrong with that. 

If you want to zero various counters on enbd (which you do NOT want to
do), you can always write "reset[a]=1" to nbdinfo. Maybe you will also
have to echo 1 to it (that downs the module in-use count), but I don't
think so. The reset command sets the openers count to zero - in fact it
sets most things internally to zero.

As to raid being locked .. can you explain what you mean?

Peter