Summary: | named pipes (FIFOs) not seen as such over NFS | ||||||
---|---|---|---|---|---|---|---|
Product: | Base System | Reporter: | Mikhail Teterin <mi> | ||||
Component: | kern | Assignee: | freebsd-bugs (Nobody) <bugs> | ||||
Status: | New --- | ||||||
Severity: | Affects Many People | CC: | jilles, pi | ||||
Priority: | --- | ||||||
Version: | 10.2-STABLE | ||||||
Hardware: | amd64 | ||||||
OS: | Any | ||||||
Attachments: |
|
Description
Mikhail Teterin
2015-11-29 06:47:31 UTC
FWIW, I do not see the same problem on another pair of servers -- both running 10.1-STABLE. Tested: nfs server 10.1p24, nfs client 10.2p7, sees: prw-rw-r-- 1 root wheel 0 Nov 29 10:12 bla Created attachment 163657 [details]
tcpdump-capture of NFS-client doing an ls
Attaching the tcpdump-capture as requested by Rick by e-mail.
Meanwhile, I tested the two computers mounting their own shares. The old:
% mount 192.168.1.8:/green /mnt
% ls -l /mnt/tmp/test
?rw-r--r-- 1 mi wheel 0 29 лис 00:05 /mnt/tmp/test
The new:
% mount 192.168.1.2:/spare /mnt
% mkfifo /spare/fifo
% ls -l /spare/fifo /mnt/fifo
prw-r--r-- 1 root wheel 0 29 лис 13:12 /mnt/fifo
prw-r--r-- 1 root wheel 0 29 лис 13:12 /spare/fifo
I then mounted the other way -- with the old system being the NFS-client -- and could not see the bug either. I think, this points conclusively at a bug in NFS-server of 10.2-STABLE #0 r289145 -- which must've been fixed by the time of
r291365.
I'll rebuild the kernel on the old server to confirm...
|