[Samba] inotify trouble

[David Bierce]

Ello --

For the past few months our Samba installation, currently at Version  
3.0.28 has filling the log. file with

[2008/04/17 13:25:30, 0] smbd/notify_inotify.c:inotify_handler(234)
  No data on inotify fd?!

This is started to happen almost weekly and it continues to output  
those lines until the log partition is full, then leaves a single  
process spiked to 100% until I kill it.  There is nothing out of the  
ordinary in the log. file prior to mass error output.  However, the  
process that is spiked (and I presume is causing the output) is always  
owned by the same user.

Their users log file is very out of the ordinary and contains lots of  
output:

rpc_api_pipe: Remote machine WK019 pipe \spoolss fnum 0x4000returned  
critical error. Error was Call timed out: server did not respond after  
10000 milliseconds
[2008/04/16 11:06:10, 0] rpc_server/ 
srv_spoolss_nt.c:srv_spoolss_replycloseprinter(161)
  srv_spoolss_replycloseprinter: reply_close_printer failed  
[WERR_SEM_TIMEOUT].
[2008/04/16 11:06:10, 0] rpc_client/cli_pipe.c:rpc_api_pipe(790)
  rpc_api_pipe: Remote machine WK019 pipe \spoolss fnum 0x4000returned  
critical error. Error was Call timed out: server did not respond after  
10000 milliseconds
[2008/04/16 11:06:10, 0] rpc_client/cli_pipe.c:rpc_api_pipe(790)
  rpc_api_pipe: Remote machine WK019 pipe \spoolss fnum 0x4000returned  
critical error. Error was Call timed out: server did not respond after  
10000 milliseconds
[2008/04/16 11:06:10, 0] rpc_client/cli_pipe.c:rpc_api_pipe(790)
  rpc_api_pipe: Remote machine WK019 pipe \spoolss fnum 0x4000returned  
critical error. Error was Call timed out: server did not respond after  
10000 milliseconds
[2008/04/16 11:09:54, 0] rpc_client/cli_pipe.c:rpc_api_pipe(790)
  rpc_api_pipe: Remote machine WK019 pipe \spoolss fnum 0x4000returned  
critical error. Error was Call timed out: server did not respond after  
10000 milliseconds
[2008/04/16 11:09:54, 0] rpc_client/cli_pipe.c:rpc_api_pipe(790)


I killed the samba process with signal -6, but the core directory is  
on the same partition as logging so it wasn't written.

Anyone else having this issue, or can think of a work around?


Thanks,
Dave