On Sun, Dec 29, 2013 at 7:26 PM, Ian Kent <raven@xxxxxxxxxx> wrote:
> On Sun, 2013-12-29 at 19:14 +0800, Ian Kent wrote:
>> On Sat, 2013-12-28 at 20:09 +0800, Dennis Lan (dlan) wrote:
>> > On Sat, Dec 28, 2013 at 10:41 AM, Ian Kent <raven@xxxxxxxxxx> wrote:
>> >
>> > >> > using commit ec267c2c406256e401540e5bc890b2e0c2ff6c35
>> > > I've updated the autofs-5.0.public branch.
>> > > Can you test it please.
>> > >
>> > > Ian
>> > >
>> > Hi Ian:
>> > Did you only fix the problem while libtirpc is disabled, as I looking
>> > at your commits?
>> > I've tested the case with libtirpc disabled, all works fine (test both
>> > nfs mount with nfsvers=3,4)
>> >
>> > but with libtirpc enabled, fail as same previous error
>>
>> I'm not able to reproduce those problems with the current source.
>> Tell me more (maybe again) about the server side.
>
what more do you want to know?
it's also a gentoo system, using kernel 3.10.9-gentoo, server nfs-utils-1.2.9
> Perhaps you could send a gdb backtrace, that might give ne a clue.
> Run gdb on the core and capture the output of "thr a a bt".
>
yes, here it is
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7ff0baf78700 (LWP 23483)]
0x0000000000000000 in ?? ()
(gdb) backtrace
#0 0x0000000000000000 in ?? ()
#1 0x00007ff0b7898d16 in clnt_vc_call (cl=0x7ff0bafed200, proc=3,
xdr_args=0x7ff0b78a1974 <xdr_rpcb>,
args_ptr=0x7ff0baf6fe50, xdr_results=0x7ff0ba0c6a10
<xdr_wrapstring>, results_ptr=0x7ff0baf6fdd8, timeout=...)
at clnt_vc.c:372
#2 0x00007ff0b67f4952 in rpc_rpcb_getport (client=0x7ff0bafed200,
parms=0x7ff0baf6fe50, timeout=...,
port=0x7ff0baf700f8) at rpc_subs.c:505
#3 0x00007ff0b67f4b7e in rpc_getport (info=0x7ff0baf6fef0,
parms=0x7ff0baf70020, client=0x7ff0bafed200,
port=0x7ff0baf700f8) at rpc_subs.c:579
#4 0x00007ff0b67f562d in rpc_portmap_getport (info=0x7ff0baf70140,
parms=0x7ff0baf70020, port=0x7ff0baf700f8)
at rpc_subs.c:927
#5 0x00007ff0b67f1c13 in get_nfs_info (logopt=1, host=0x7ff0bafdd600,
pm_info=0x7ff0baf70140,
rpc_info=0x7ff0baf700e0, proto=6, version=32, port=0) at replicated.c:535
#6 0x00007ff0b67f2303 in get_vers_and_cost (logopt=1,
host=0x7ff0bafdd600, version=35, port=-1)
at replicated.c:686
#7 0x00007ff0b67f2ba0 in prune_host_list (logopt=1,
list=0x7ff0baf70348, vers=35, port=-1) at replicated.c:910
#8 0x00007ff0b67f0319 in mount_mount (ap=0x7ff0bafd9600,
root=0x7ff0bafeaeb0 "/mnt/misc",
name=0x7ff0baf71500 "mkynfs1", name_len=7, what=0x7ff0baf714d0
"192.168.90.12:/home/dennis",
fstype=0x7ff0b6a3ff8e "nfs", options=0x7ff0baf71520
"ro,soft,intr,nfsvers=3", context=0x7ff0bafe2e00)
at mount_nfs.c:233
#9 0x00007ff0b6a26029 in sun_mount (ap=0x7ff0bafd9600,
root=0x7ff0bafeaeb0 "/mnt/misc",
name=0x7ff0baf75940 "mkynfs1", namelen=7, loc=0x7ff0bafe2780
"192.168.90.12:/home/dennis", loclen=26,
options=0x7ff0baf71520 "ro,soft,intr,nfsvers=3",
ctxt=0x7ff0bafe2de0) at parse_sun.c:657
#10 0x00007ff0b6a291c8 in parse_mount (ap=0x7ff0bafd9600,
name=0x7ff0baf75940 "mkynfs1", name_len=7,
mapent=0x7ff0baf71930
"-ro,soft,intr,nfsvers=3\t\t192.168.90.12:/home/dennis",
context=0x7ff0bafe2de0)
at parse_sun.c:1595
#11 0x00007ff0b6c5619b in lookup_mount (ap=0x7ff0bafd9600,
name=0x7ff0baf76eb0 "mkynfs1", name_len=7,
context=0x7ff0bafe2e20) at lookup_file.c:1142
#12 0x00007ff0bb011d52 in do_lookup_mount (ap=0x7ff0bafd9600,
map=0x7ff0bafddd80, name=0x7ff0baf76eb0 "mkynfs1",
name_len=7) at lookup.c:688
#13 0x00007ff0bb011f09 in lookup_name_file_source_instance
(ap=0x7ff0bafd9600, map=0x7ff0bafdda80,
name=0x7ff0baf76eb0 "mkynfs1", name_len=7) at lookup.c:728
#14 0x00007ff0bb01269c in lookup_nss_mount (ap=0x7ff0bafd9600,
source=0x0, name=0x7ff0baf76eb0 "mkynfs1",
name_len=7) at lookup.c:920
#15 0x00007ff0bb007c2e in do_mount_indirect (arg=0x7ff0bae75000) at
indirect.c:776
#16 0x00007ff0babbd013 in start_thread () from /lib64/libpthread.so.0
#17 0x00007ff0ba0989dd in clone () from /lib64/libc.so.6
>>
>> Commit ec267c2c deals with a libtirpc failure that would result in the
>> "no hosts found" message and I'm not sure you had commit 7f7b730e either
>> which also caused failures.
>>
I'm testing autofs-5.0.public branch (commit:
3d4c9d156c065afd0b38491cec4367b1ecbae24a)
previous report based on (e7ada505a83871922a588d50da06d02000325f1a)
so, yes, they all include ec267c2c and 7f7b730e
0
>> Ian
>>
>
>
On Sun, Dec 29, 2013 at 7:36 PM, Ian Kent <raven@xxxxxxxxxx> wrote:
> On Fri, 2013-12-27 at 15:30 +0800, Dennis Lan (dlan) wrote:
>> sorry, I was not accurate about the failure cases.
>>
>> using commit ec267c2c406256e401540e5bc890b2e0c2ff6c35
>>
>> snip from file /etc/autofs/auto.misc
>> mkynfs -ro,soft,intr,nfsvers=3 192.168.90.12:/home/dennis
>> mkydns -ro,soft,intr,nfsvers=4 mickeycn:/home/dennis
>>
>> with libtirpc enabled (all other features disabled),[1]
>> "ls /mnt/misc/mkynfs" lead to automount segfault
>> "ls /mnt/misc/mkydns" automount fail (no crash), but report "no hosts available"
>>
>> with libtirpc disabled (all other features disabled)[2]
>> "ls /mnt/misc/mkynfs" automount fail (no crash), report "no hosts available"
>> "ls /mnt/misc/mkydns" works fine
>
> What exactly does "all other features disabled" mean from a configure
> POV. I always use --disable-mount-locking and --enable-ignore-busy,
> always.
>
> Maybe that's something to do with the problem your seeing .....
>
> Ian
--
To unsubscribe from this list: send the line "unsubscribe autofs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
