Bug 984824
Summary: | [rpc.mountd] -n|--no-tcp option cannot work with -N 4 | |||
---|---|---|---|---|
Product: | Red Hat Enterprise Linux 6 | Reporter: | JianHong Yin <jiyin> | |
Component: | nfs-utils | Assignee: | Steve Dickson <steved> | |
Status: | CLOSED WONTFIX | QA Contact: | JianHong Yin <jiyin> | |
Severity: | high | Docs Contact: | ||
Priority: | high | |||
Version: | 6.4 | CC: | jiyin, rwheeler, steved, swhiteho | |
Target Milestone: | rc | Keywords: | Reopened, TestBlocker | |
Target Release: | --- | |||
Hardware: | All | |||
OS: | Linux | |||
Whiteboard: | ||||
Fixed In Version: | Doc Type: | Bug Fix | ||
Doc Text: | Story Points: | --- | ||
Clone Of: | ||||
: | 1013468 1127955 (view as bug list) | Environment: | ||
Last Closed: | 2014-08-07 14:47:01 UTC | Type: | Bug | |
Regression: | --- | Mount Type: | --- | |
Documentation: | --- | CRM: | ||
Verified Versions: | Category: | --- | ||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | ||
Cloudforms Team: | --- | Target Upstream Version: | ||
Embargoed: | ||||
Bug Depends On: | ||||
Bug Blocks: | 859078, 1011909, 1013468, 1127955, 1127957 |
Description
JianHong Yin
2013-07-16 06:48:24 UTC
in rhel6.0 release version has the same problem. the package version nfs-utils-1.2.3-36.el6 and in a rhel6.0 alpha distro env. rpc.mountd -n works fine. package version nfs-utils-1.2.1-10.el6 so the bug introduced interval (1.2.1-10, 1.2.3-36] RHEL6.5 has the same issue The reason --no-tcp does not work without turning off v4 (aka -N 4) is because tcp is *required* by v4... If TCP could be turned off without turning off v4, then that was bug... (In reply to Steve Dickson from comment #5) > The reason --no-tcp does not work without turning off v4 (aka -N 4) is > because > tcp is *required* by v4... If TCP could be turned off without turning > off v4, then that was bug... sorry, my mistake for description; yes with -n -N 4 rpc.mountd still can not turn off TCP. - And it cause bz859078 test case can not work. Hi Steve in RHEL7 has the same problem: bz1013468 please fix them together. (In reply to Yin.JianHong from comment #6) > (In reply to Steve Dickson from comment #5) > > The reason --no-tcp does not work without turning off v4 (aka -N 4) is > > because > > tcp is *required* by v4... If TCP could be turned off without turning > > off v4, then that was bug... > > sorry, my mistake for description; > > yes with -n -N 4 rpc.mountd still can not turn off TCP. > - And it cause bz859078 test case can not work. I'm a bit confused there... The title of this bz is [rpc.mountd] -n|--no-tcp option cannot work which is not the case as you have proven. You also need the -N 4 to disable tcp, which is how it worked since RHEL6.0. Point being this is not a regression. So why is this a Test Blocker?? (In reply to Steve Dickson from comment #8) > (In reply to Yin.JianHong from comment #6) > > (In reply to Steve Dickson from comment #5) > > > The reason --no-tcp does not work without turning off v4 (aka -N 4) is > > > because > > > tcp is *required* by v4... If TCP could be turned off without turning > > > off v4, then that was bug... > > > > sorry, my mistake for description; > > > > yes with -n -N 4 rpc.mountd still can not turn off TCP. > > - And it cause bz859078 test case can not work. > I'm a bit confused there... The title of this bz is > [rpc.mountd] -n|--no-tcp option cannot work Sorry for the confuse,, I changed the title. > > which is not the case as you have proven. You also need > the -N 4 to disable tcp, which is how it worked since > RHEL6.0. Point being this is not a regression. > > So why is this a Test Blocker?? because another bug's reproducer need "-n -N 4" disable tcp port. I think that you need to fix your older test - that is a really old BZ, right? (In reply to Ric Wheeler from comment #10) > I think that you need to fix your older test - that is a really old BZ, > right? Is this reply for me? or steve? If for me, the answer is no, that is a new test for: mount.nfs negotiates suitable version with nfs server. And I think if just specify -n option, rpc.mountd should output warning msg to user. or add doc description in man page. This is listed as a release blocker for RHEL6.5. I don't see this as a critical issue and clearly it is not a blocker if this is just worked around with a man page update. I am going to push it back to RHEL6.6, thanks! (In reply to Ric Wheeler from comment #12) > This is listed as a release blocker for RHEL6.5. I don't see this as a > critical issue and clearly it is not a blocker if this is just worked around > with a man page update. No no no, update man page is an other question; The bug in this bugzilla, is: rpc.mountd -n -N 4 can not work as expected, can not turn off the tcp port > > I am going to push it back to RHEL6.6, thanks! I'm not sure I understand what the issue is here. NFSv4 doesn't support udp, so setting the version of NFS to 4, and saying "don't use tcp" doesn't make sense as what other transport can it use? (In reply to Steve Whitehouse from comment #15) > I'm not sure I understand what the issue is here. NFSv4 doesn't support udp, > so setting the version of NFS to 4, and saying "don't use tcp" doesn't make no, the options "-N 4 -n" means "disable nfs4 and disable tcp" ------------------------------------------- Info of the -N option in man rpc.mountd(8) -N or --no-nfs-version This option can be used to request that rpc.mountd do not offer certain versions of NFS. The current version of rpc.mountd can support both NFS version 2, 3 and 4. If the either one of these version should not be offered, rpc.mountd must be invoked with the option --no-nfs-version <vers> . ------------------------------------------- > sense as what other transport can it use? :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: :: [ LOG ] :: /CoreOS/nfs-utils/commands/utils/rpc-mountd-Test-rpc.mountd -n --no-tcp :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: [23:29:21 root@ ~~]# rpc.mountd -n -N 4 :: [ PASS ] :: Running 'rpc.mountd -n -N 4' (Expected 0, got 0) -------------------------------------------------------------------------------- [23:29:22 root@ ~~]# ps axf | grep -v grep | grep "rpc.mountd -n" 4702 ? Ss 0:00 rpc.mountd -n -N 4 :: [ PASS ] :: Running 'ps axf | grep -v grep | grep "rpc.mountd -n"' (Expected 0, got 0) -------------------------------------------------------------------------------- [23:29:22 root@ ~~]# rpcinfo -p | grep -w mountd | grep -w tcp 100005 1 tcp 41602 mountd 100005 2 tcp 56913 mountd 100005 3 tcp 48060 mountd :: [ FAIL ] :: should no tcp in rpcbind, rhel7(bz984824) (Expected 1, got 0) -------------------------------------------------------------------------------- [23:29:22 root@ ~~]# pkill rpc.mountd -------------------------------------------------------------------------------- [23:29:22 root@ ~~]# rpc.mountd --no-tcp -N 4 :: [ PASS ] :: Running 'rpc.mountd --no-tcp -N 4' (Expected 0, got 0) -------------------------------------------------------------------------------- [23:29:23 root@ ~~]# ps axf | grep -v grep | grep "rpc.mountd --no-tcp" 4776 ? Ss 0:00 rpc.mountd --no-tcp -N 4 :: [ PASS ] :: Running 'ps axf | grep -v grep | grep "rpc.mountd --no-tcp"' (Expected 0, got 0) -------------------------------------------------------------------------------- [23:29:23 root@ ~~]# rpcinfo -p | grep -w mountd | grep -w tcp 100005 1 tcp 45682 mountd 100005 2 tcp 52036 mountd 100005 3 tcp 38612 mountd :: [ FAIL ] :: should no tcp in rpcbind, rhel7(bz984824) (Expected 1, got 0) -------------------------------------------------------------------------------- [23:29:23 root@ ~~]# pkill rpc.mountd -------------------------------------------------------------------------------- Ok, sorry. I see now. Thanks for confirming, that does sound like this needs to be fixed in that case. (In reply to Yin.JianHong from comment #13) > (In reply to Ric Wheeler from comment #12) > > This is listed as a release blocker for RHEL6.5. I don't see this as a > > critical issue and clearly it is not a blocker if this is just worked around > > with a man page update. > No no no, update man page is an other question; > The bug in this bugzilla, is: > rpc.mountd -n -N 4 can not work as expected, can not turn off the tcp > port It is true that -n no longer turns off tcp listeners. It happen when we switch over from using glibc to libtirpc which was a number of years ago. In this day and age, the last thing customers need to do is turn off tcp. Plus the only thing that has notice anomaly is our QE scripts, meaning no customer complaints at all... So I suggest we close this bz as WONTFIX and open another bz to remove the -n from the man page... |