bug-coreutils
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: du make check failure on Solaris 5.5 and 5.8 in coreutils-4.5.9


From: Vin Shelton
Subject: Re: du make check failure on Solaris 5.5 and 5.8 in coreutils-4.5.9
Date: Tue, 11 Mar 2003 14:12:14 -0500

Jim Meyering <address@hidden> writes:

> Vin Shelton <address@hidden> wrote:
>> Jim,
>>
>> Jim Meyering <address@hidden> writes:
>>> Vin Shelton <address@hidden> wrote:
>>>> In coreutils-4.5.9 on the Sparc Solaris machines at work, I'm getting
>>>> a 'make check' failure in the basic du test.  Here is the failure
>>>> under Solaris-5.5; as you can see the basic test fails:
>>>
>>> Hi Vin!
>>>
>>> Thank you for the report!
>>> Here's a patch:
>>>
>>>   http://mail.gnu.org/archive/html/bug-coreutils/2003-03/msg00046.html
>>>
>> Thanks for the speedy response, but the proposed patch didn't work.
>
> I've looked at your actual results this time :-)
> The problem appears to be that your filesystem (BTW, which type is it?)
> requires zero blocks for files like a/b/c that are small enough.
> I think I've fixed the test script for next time by requiring
> that the blocksize and block-count match for files as well as for
> directories.  So on your system, the new du/basic test would be skipped.
> But maybe I can do better:
>
> Would you please find out how big a file has to be in order to get a
> non-zero block count?  Then I'll change the test so that the files
> it uses are at least that big.

This is apparently an artifact of the Network Appliance (NFS-mounted)
filesystem. Here are the results of:

    for i in {64..66}; do
    echo "\n$i"
    dd if=/dev/zero of=testfile count=1 bs=$i
    ls -l testfile
    /usr/bin/du -a testfile;du -a -B 512 testfile
    done

64
1+0 records in
1+0 records out
-rw-r--r--    1 shelton  vobgrp         64 Mar 11 14:05 testfile
0       testfile
0       testfile

65
1+0 records in
1+0 records out
-rw-r--r--    1 shelton  vobgrp         65 Mar 11 14:05 testfile
8       testfile
8       testfile

66
1+0 records in
1+0 records out
-rw-r--r--    1 shelton  vobgrp         66 Mar 11 14:05 testfile
8       testfile
8       testfile


When I run that same test on a local fs, here's what I see:

64
1+0 records in
1+0 records out
-rw-r--r--    1 shelton  vobgrp         64 Mar 11 14:07 testfile
16      testfile
16      testfile

65
1+0 records in
1+0 records out
-rw-r--r--    1 shelton  vobgrp         65 Mar 11 14:07 testfile
16      testfile
16      testfile

66
1+0 records in
1+0 records out
-rw-r--r--    1 shelton  vobgrp         66 Mar 11 14:07 testfile
16      testfile
16      testfile


HTH,
  Vin




reply via email to

[Prev in Thread] Current Thread [Next in Thread]