View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0006213||CentOS-5||nfs-utils||public||2013-01-23 08:24||2013-04-18 20:02|
|Reporter||Andreas Baumann||Assigned To|
|Summary||0006213: NFS mounts don't show all directory entries|
|Description||'ls' on a NFS client shows only partial listings. The same directory looks ok|
when doing an 'ls' on the NFS server.
I strongly suspect the NFSv4 readdir glitch at work as mentioned in:
I get the same behaviour as mentioned in:
|Tags||No tags attached.|
I just run into a similar problem. Don't know, if it is the same, as i don't know if my old LTS mounts with v4, but with Centos kernel 2.6.18-348 i immediately cannot mount from LTSP-4 clients (a old kernel 2.4.26) to the 5.9 server. A ls command returns emtpy, X11 looking for modules to load fail with "doesn't exist".
If i do a "cat filename" or "more filename", the file is accessed. Only reading the dir seems not to workl.
Reverting to kernel 2.6.18-308.24.1.el5 brings back the old functionality.
Looks Centos 6.2 had the same problem, 6.3 is fine:
Maybe this has the same root cause?
|May be related to (or a duplicate of) bug #6241.|
|kernel-2.6.18-348.4.1.el5 has been released. Does the problem still exist in this version?|
|Sadly I replaced Centos 5 on the NAS and can't test it easily at the moment.. :-(|
Will try and report as soon as possible.
Have a development system which currently updates to .4.1 but cannot be rebooted until night (MEST).
|Unfortunately i have to report, that the new kernel didn't change anything. After mounting NFS vol exported from a 5.9 system, a ls (readdir) doesn't show any files. Direct access by (known) name still possible.|
I am still not quite sure if the problem you are seeing is the same as the OP's. Hope Andreas Baumann can test the latest kernel and see if the original issue is resolved.
I was seeing the same problem as the OP and kernel-2.6.18-348.4.1.el5 appears to have resolved it. ls on a directory containing a large number of files returns only a partial listing when the NFS server is booted with any previous 5.9 kernel (2.6.18-348 through 2.6.18-348.3.1).
Server: CentOS 5.9 32-bit
Client: CentOS 6.4 32-bit
Thanks for the report. That confirms that the following description in the kernel release note applies to the current bug.
"Previously, the NFSv3 server incorrectly converted 64-bit cookies to
32-bit. Consequently, the cookies became invalid, which affected all file
system operations depending on these cookies, such as the READDIR operation
that is used to read entries from a directory. This led to various
problems, such as exported directories being empty or displayed
incorrectly, or an endless loop of the READDIRPLUS procedure which could
potentially cause a buffer overflow. This update modifies knfsd code so
that 64-bit cookies are now handled correctly and all file system
operations work as expected. (BZ#924087)"
|2013-01-23 08:24||Andreas Baumann||New Issue|
|2013-01-31 11:42||rkollien||Note Added: 0016368|
|2013-01-31 12:05||Andreas Baumann||Note Added: 0016369|
|2013-02-11 19:32||toracat||Relationship added||related to 0006241|
|2013-02-11 19:33||toracat||Note Added: 0016480|
|2013-04-16 20:09||toracat||Relationship added||related to 0005496|
|2013-04-17 12:19||toracat||Note Added: 0017224|
|2013-04-17 12:19||toracat||Status||new => feedback|
|2013-04-17 12:28||Andreas Baumann||Note Added: 0017226|
|2013-04-17 12:28||Andreas Baumann||Status||feedback => assigned|
|2013-04-17 12:35||rkollien||Note Added: 0017227|
|2013-04-18 05:29||rkollien||Note Added: 0017229|
|2013-04-18 15:29||toracat||Note Added: 0017231|
|2013-04-18 19:40||odimond||Note Added: 0017233|
|2013-04-18 20:02||toracat||Note Added: 0017234|