Messages in this thread | | | Date | Thu, 20 Nov 1997 23:00:53 -0500 | From | Bill Hawes <> | Subject | Re: patch for 2.1.65 smbfs |
| |
Gordon Chaffee wrote:
> I'm seeing very poor behavior when listing directories with both a clean > 2.1.65 and with your patch applied. > > zeego chaffee[524] time ls -l /hosts/feba/f/exceed.nt/ > [output deleted] > 0.04user 0.06system 0:48.77elapsed 0%CPU (0avgtext+0avgdata 0maxresident)k > 0inputs+0outputs (0major+0minor)pagefaults 0swaps > > With no options to ls, it goes at a reasonable speed. > > zeego chaffee[525] time ls /hosts/feba/f/exceed.nt/ > 0.02user 0.00system 0:00.90elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k > 0inputs+0outputs (0major+0minor)pagefaults 0swaps
Hmm, I'm not sure what can be done to speed up ls -l -- it's doing a getattr on the files, so it generates a lot of message traffic.
> > I'm also still seeing the timeout problems. Things will be working fine > for awhile, and then I get these message in my log file. > > Nov 20 13:20:16 zeego kernel: smb_get_length: recv error = 512 > Nov 20 13:20:16 zeego kernel: smb_request: result -512, setting invalid > Nov 20 13:20:16 zeego kernel: Busy Inodes for Device 00:03: > Nov 20 13:20:16 zeego kernel: inode 9282: count=1 state=0 > Nov 20 13:20:16 zeego kernel: inode 9281: count=1 state=0 > Nov 20 13:20:16 zeego kernel: inode 7863: count=1 state=0 > Nov 20 13:20:28 zeego kernel: smb_retry: timed out, try again later > Nov 20 13:20:28 zeego kernel: smb_lookup: find //EMACS-19.34 failed, error=-5
It seems like something must be wrong with the samba server process, as it normally should respond instantly with a new socket. I'm not too clear on what happens on the samba side when you need a new connection -- smbfs just sends the server a signal and waits for an ioctl call.
Any way you could turn on debug logging on the samba side and see what happens during these events?
Regards, Bill
| |