weirdest problem, from a RH7.2 (base or + everything) samba client running
eclipse utilizing SMB mount.
Apr 11 13:43:54 testsmbbug kernel: smb_get_length: recv error = 512
Apr 11 13:43:54 testsmbbug kernel: smb_request: result -512, setting invalid
Apr 11 13:43:54 testsmbbug kernel: smb_retry: caught signal
Apr 11 13:43:54 testsmbbug kernel: smb_retry: caught signal
Apr 11 13:43:55 testsmbbug kernel: smb_retry: successful, new pid=18163,
generation=8
Apr 11 13:43:57 testsmbbug kernel: smb_get_length: recv error = 512
Apr 11 13:43:57 testsmbbug kernel: smb_request: result -512, setting invalid
Apr 11 13:43:57 testsmbbug kernel: smb_retry: caught signal
Apr 11 13:43:57 testsmbbug kernel: smb_retry: caught signal
Apr 11 16:28:23 testsmbbug kernel: smb_get_length: recv error = 512
Apr 11 16:28:23 testsmbbug kernel: smb_request: result -512, setting invalid
Apr 11 16:28:23 testsmbbug kernel: smb_retry: caught signal
Apr 11 16:28:23 testsmbbug kernel: smb_retry: caught signal
Apr 11 16:28:24 testsmbbug kernel: smb_retry: successful, new pid=18163,
generation=10
Apr 11 16:28:32 testsmbbug kernel: smb_get_length: recv error = 512
Apr 11 16:28:32 testsmbbug kernel: smb_request: result -512, setting invalid
Apr 11 16:28:32 testsmbbug kernel: smb_retry: caught signal
We had only seen it with with eclipse on linux, then we found that kghostview
will do it when doing a zoom (in or out), every time, when viewing a pdf over
an smb mount..
Initially identified on rh7.2+ updates on P4 1.6GHz Asus P4B. But since
identified on P3 800MHz. all against samba 2.2.2 and 2.2.3a and Window's
server shares. I have win98 eclipse clients hitting it fine.
applied every update available, including latest glib update and kernels from
RH. same thing.
updated the kernels to 2.4.18 with/without pre5/6 ac3 etc. at least smb_retry
reconnects now.
applied 00-02 patches from Urban Widmark's smbfs site and patched smbmount,
recompiled and reinstalled. no change.
Thought it might be related to 3com nic (all effected systems had varieties of
3c905c(x)'s). installed an IBM pci TR adapter, same thing. damn, though i
was on to something.
Did a network trace (tcpdump) on the client and there is NO traffic to the
smb server before or during the kghostview zoom hang (hangs for 10 seconds
then generates the above errors) until after the errors are posted to the
/var/log/messages, then what looks like normal connect/re-connect traffic
(again it's after the fact).
debug on smbmount shows nothing after the mount occurs. You can access the
mounted FS fine from the command line, vi, etc. Something in the way these
apps access the FS is making smbfs think it has lost the connection (but
retry works - after a few seconds - under 2.4.18). I've been able to
generate the same errors doing rpm updates over an smb mount occasionaly on
these systems as well (found a post or two regarding simliar problems with
remote gdb databases - applying smbfs patches didn't help)
this looks like a timing thing, but I can't find it. smbfs or somewhere
upstream. It appears to only happen on some faster processors. I have a
500mhz PIII 7.2 +updates that doesn't and have tried it from one of the
samba
servers to the other (SMP PIII 800's) and it didn't have a problem. The
network is clean and idling, all error counters in ifconfig and
/proc/net/snmp are 0 on all systems tested.
any Ideas?
--
Brian Spears
Dir I/t
Konica Business Technologies