Bug 1218045
| Summary: | Similar but unrelated NFS exports block proper mounting of "parent" mount point | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | Ian Kent <ikent> |
| Component: | autofs | Assignee: | Ian Kent <ikent> |
| Status: | CLOSED ERRATA | QA Contact: | Yongcheng Yang <yoyang> |
| Severity: | unspecified | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 7.1 | CC: | eguan, fs-qe, ikent, pere |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | (autofs-5.0.7-50 | Doc Type: | Bug Fix |
| Doc Text: |
Cause: The function cache_get_offset() which returns offsets to be mounted that are within the current current subtree for the mount-as-you-go functionality was not returning the correct map entries.
Consequence: This prevented mounts higher up in the tree from being mounted because nested offset triggers created within these mounts would make them look like they weren't mount points to the autofs kernel module.
Fix: The offset selection function was modified to return the correct offsets for the given level of the mount tree.
Result: Mount trees are now mounted "as-you-go" in the correct order so that mounts higher up in the tree are no longer prevented from being mounted.
|
Story Points: | --- |
| Clone Of: | 1203139 | Environment: | |
| Last Closed: | 2015-11-19 13:00:56 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: | 1203139 | ||
| Bug Blocks: | |||
|
Comment 2
XuWang
2015-08-28 11:38:51 UTC
Tiny issues, please NOTICE:
1.the following change seems not right, but no affect for usage:
if (pstart[len] != '/' ||
strlen(pstart) != len ||
^^^^^^^^^^^^^^^^^^^^ this is the new condition added, but if the multi map is like "/a, /a/b, /a/b/c", and for the first time, the ofset will be /a, and in this function, should ignore /a/b and /a/b/c, but with the "strlen(pstart) != len", we can't ignore, only wait for next calling for "cache_get_offset".
strncmp(offset, pstart, len
2. if the nfs server is local relative to autofs, the the symbol link will be used, and a dead lock will happen:
/a/b will be mounted autofs, and further accessing to /a/b for c will hung.
But in my idea, this is the supposed way.
(In reply to XuWang from comment #3) > Tiny issues, please NOTICE: Thanks for the comments. This particular bit of code is complicated so it is likely not quite right. This function isn't meant to get all offsets on each call, it's meant to get all offsets up to a nesting point boundary. This is crux of the "mount as you go" functionality and is what makes it more complicated than I would like. I'll check it out as soon as I get time. All the testing I did worked OK and it appears yours has too so I don't think this is urgent and I have some other high priority work at the moment. > > 1.the following change seems not right, but no affect for usage: > > if (pstart[len] != '/' || > strlen(pstart) != len || > ^^^^^^^^^^^^^^^^^^^^ this is the new condition added, but if the multi > map is like "/a, /a/b, /a/b/c", and for the first time, the ofset will be > /a, and in this function, should ignore /a/b and /a/b/c, but with the > "strlen(pstart) != len", we can't ignore, only wait for next calling for > "cache_get_offset". > > strncmp(offset, pstart, len > > 2. if the nfs server is local relative to autofs, the the symbol link will > be used, and a dead lock will happen: > /a/b will be mounted autofs, and further accessing to /a/b for c will hung. > But in my idea, this is the supposed way. Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHSA-2015-2417.html |