Trouble setting up mogileFS

Paul Hurley paul at 1100tech.com
Mon Jan 30 01:23:24 UTC 2006


Oh yeah I guess I forgot that critical part to the story, but glad you 
recognized it. I did have errors when installing Linux::AIO and forced 
it. I did install the correct package though, but here is the error I got.
(I'm running CentOS 4.2. perl 5.8.5, pretty much out of the box install 
in a vmware)


# make test
gcc -c   -D_REENTRANT -D_GNU_SOURCE -DDEBUGGING -fno-strict-aliasing 
-pipe -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 
-I/usr/include/gdbm -O2 -g -pipe -m32 -march=i386 -mtune=pentium4   
-DVERSION=\"1.8\" -DXS_VERSION=\"1.8\"  
"-I/usr/lib/perl5/5.8.5/i386-linux-thread-multi/CORE"   AIO.c
Running Mkbootstrap for Linux::AIO ()
chmod 644 AIO.bs
rm -f blib/arch/auto/Linux/AIO/AIO.so
gcc  -shared -L/usr/local/lib AIO.o  -o blib/arch/auto/Linux/AIO/AIO.so
chmod 755 blib/arch/auto/Linux/AIO/AIO.so
cp AIO.bs blib/arch/auto/Linux/AIO/AIO.bs
chmod 644 blib/arch/auto/Linux/AIO/AIO.bs
PERL_DL_NONLAZY=1 /usr/bin/perl "-MExtUtils::Command::MM" "-e" 
"test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
t/00_load......ok
t/01_stat......ok
t/02_read......ok
t/03_errors....NOK 1# Failed test 1 in t/03_errors.t at line 21
#  t/03_errors.t line 21 is:     ok($_[0] < 0 && $! == ENOENT);
t/03_errors....NOK 7# Failed test 7 in t/03_errors.t at line 40
#  t/03_errors.t line 40 is:     ok($! == ENOENT);
t/03_errors....NOK 10# Failed test 10 in t/03_errors.t at line 49
#  t/03_errors.t line 49 is:     ok($! == EBADF);
t/03_errors....FAILED tests 1, 7, 10
        Failed 3/10 tests, 70.00% okay
Failed Test   Stat Wstat Total Fail  Failed  List of Failed
-------------------------------------------------------------------------------
t/03_errors.t               10    3  30.00%  1 7 10
Failed 1/4 test scripts, 75.00% okay. 3/26 subtests failed, 88.46% okay.
make: *** [test_dynamic] Error 255


I tried IO::AIO as you suggested, here was the error during that test 
making IO::AIO version 1.71:

Running make test
PERL_DL_NONLAZY=1 /usr/bin/perl "-MExtUtils::Command::MM" "-e" 
"test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
t/00_load.......ok
t/01_stat.......ok
t/02_read.......dubious
        Test returned status 255 (wstat -1, 0xffffffff)
        after all the subtests completed successfully
t/03_errors.....ok
t/04_fork.......ok
t/05_readdir....ok
Failed Test Stat Wstat Total Fail  Failed  List of Failed
-------------------------------------------------------------------------------
t/02_read.t  255    -1     2    0   0.00%  ??
Failed 1/6 test scripts, 83.33% okay. 0/44 subtests failed, 100.00% okay.
make: *** [test_dynamic] Error 255
  /usr/bin/make test -- NOT OK





Brad Fitzpatrick wrote:

>Are you using Linux::AIO or IO::AIO?  My guess is you installed Linux::AIO
>but the test suite didn't pass.  Go download it,
>
>http://search.cpan.org/CPAN/authors/id/M/ML/MLEHMANN/Linux-AIO-1.8.tar.gz
>
>and then:
>
>$ perl Makefile.PL
>$ make test
>
>What does it say?
>
>If it fails, could you also try IO::AIO and let me know what that test
>suite says.  I may have to change the default preference to use IO::AIO
>instead of Linux::AIO.
>
>http://search.cpan.org/CPAN/authors/id/M/ML/MLEHMANN/IO-AIO-1.71.tar.gz
>
>(Anybody reading this that wants to test both, I'm interested in hearing
>your results....)
>
>- Brad
>
>
>On Sun, 29 Jan 2006, Arnoud Vermeer wrote:
>
>  
>
>>I fixed the problem by manually creating the dir dev[456] in /var/mogdata.
>>It seems like there is a problem with the mogstored demon (I think?)
>>http://lists.danga.com/pipermail/mogilefs/2006-January/000244.html is having
>>the same problem as I am.
>>
>>I hope this problem can get fixed.
>>
>>Thanks,
>>
>>Arnoud Vermeer
>>
>>On 1/29/06, Arnoud Vermeer <a.vermeer at freshway.biz> wrote:
>>    
>>
>>>Thanks for your replies.
>>>
>>>Regarding Joshua:
>>>mfs1:# df -l
>>>Filesystem           1K-blocks      Used Available Use% Mounted on
>>>/dev/hda1             37776652    555432  35302260   2% /
>>>tmpfs                   124472         0    124472   0% /dev/shm
>>>
>>>mfs2:# df -l
>>>Filesystem           1K-blocks      Used Available Use% Mounted on
>>>/dev/hda1             37136300    501176  34748696   2% /
>>>tmpfs                   225764         0    225764   0% /dev/shm
>>>
>>>mfs3:# df -l
>>>Filesystem           1K-blocks      Used Available Use% Mounted on
>>>/dev/hda1             37744520    510772  35316392   2% /
>>>tmpfs                   128520         0    128520   0% /dev/shm
>>>
>>>I didn't map any device to the devicenames in mogilefs, should I ? and
>>>how?
>>>
>>>Regarding brad:
>>>
>>>telnet localhost 7500
>>>Trying 127.0.0.1...
>>>Connected to localhost.localdomain.
>>>Escape character is '^]'.
>>>GET http://81.171.121.30:7500/dev4/usage
>>>
>>>doesn't return anything..
>>>
>>>This is my config:
>>>cat /etc/mogilefs/mogstored.conf
>>>httplisten = 0.0.0.0:7500
>>>docroot = /var/mogdata
>>>
>>>--
>>>
>>>cat /etc/mogilefs/mogilefsd.conf
>>>daemonize = 1
>>>db_dsn = DBI:mysql:mogilefs:host=111.111.111.43
>>>db_user = mogilefs
>>>db_pass = PASSWD
>>>conf_port = 7001
>>>listener_jobs = 10
>>>delete_jobs = 1
>>>replicate_jobs = 5
>>>mog_root = /mnt/mogilefs
>>>reaper_jobs = 1
>>># debug = 3
>>>
>>>any suggestions?
>>>
>>>Thanks,
>>>
>>>Arnoud
>>>
>>>On 1/29/06, Brad Fitzpatrick <brad at danga.com> wrote:
>>>      
>>>
>>>>You have some port numbers wrong somewhere... the tracker doesn't speak
>>>>HTTP.
>>>>
>>>>- Brad
>>>>
>>>>
>>>>On Sat, 28 Jan 2006, Arnoud Vermeer wrote:
>>>>
>>>>        
>>>>
>>>>>Hi guys,
>>>>>
>>>>>I'm trying to setup a small test cluster of MogileFS, but I'm running
>>>>>          
>>>>>
>>>>into a
>>>>        
>>>>
>>>>>problem.
>>>>>I've got 3 desktop PC's in the datacenter running Debian Sarge
>>>>>          
>>>>>
>>>>(installed
>>>>        
>>>>
>>>>>via netinstall)
>>>>>2 x 1200 Mhz intel with a single PATA 40 gb HD
>>>>>and a AMD 2400+ also with a PATA 40 gb HD.
>>>>>
>>>>>I've installed mogileFS just fine, first the .tar.gz release from the
>>>>>          
>>>>>
>>>>site,
>>>>        
>>>>
>>>>>but that one was running into problems with adding disks.
>>>>>So I grabbed the latest CVS (28-jan-2006), but this one seems to run
>>>>>          
>>>>>
>>>>into
>>>>        
>>>>
>>>>>the same problems.
>>>>>
>>>>>mfs1 (1200Mhz+40GB HD) runs Perlbal, mogilefsd and mogstored
>>>>>mfs2 (2400Mhz+40GB HD) runs Perlbal, mogilefsd and mogstored
>>>>>mfs3 (1200Mhz+40GB HD) just runs mogstored
>>>>>
>>>>>I've added the hosts and disks via the mogadm commandline, but when I
>>>>>          
>>>>>
>>>>run a:
>>>>        
>>>>
>>>>>mogadm -c /etc/mogilefs/mogadm.conf check
>>>>>
>>>>>I get the following result:
>>>>>
>>>>>mogadm -c /etc/mogilefs/mogadm.conf check (IP's and domain removed..)
>>>>>Checking trackers...
>>>>>        111.111.111.30:7001 ... OK
>>>>>        111.111.111.31:7001 ... OK
>>>>>
>>>>>Checking hosts...
>>>>>        [ 1] mfs1.domain ... OK
>>>>>        [ 2] mfs2.domain ... OK
>>>>>        [ 3] mfs3.domain ... OK
>>>>>
>>>>>Checking devices...
>>>>>        host device            size(G)    used(G)    free(G)   use%
>>>>>        ---- --------------- ---------- ---------- ---------- ------
>>>>>        [ 1] dev4      REQUEST FAILURE
>>>>>        [ 2] dev3      REQUEST FAILURE
>>>>>        [ 3] dev5      REQUEST FAILURE
>>>>>Illegal division by zero at /usr/bin/mogadm line 165,
>>>>><Sock_81.171.121.31:7001> line 2.
>>>>>
>>>>>So I looked into the source and found the command at line 165, and
>>>>>          
>>>>>
>>>>gave this
>>>>        
>>>>
>>>>>via telnet onto the the tracker:
>>>>>
>>>>>telnet localhost 7001
>>>>>Trying 127.0.0.1...
>>>>>Connected to localhost.localdomain.
>>>>>Escape character is '^]'.
>>>>>GET http://81.171.121.30:7500/dev4/usage
>>>>>ERR unknown_command Unknown+server+command
>>>>>
>>>>>So now I'm stumped.
>>>>>
>>>>>Does anybody have a clue?
>>>>>
>>>>>Thanks,
>>>>>
>>>>>Arnoud Vermeer
>>>>>FreshWay Innovations
>>>>>
>>>>>          
>>>>>
>>>      
>>>


More information about the mogilefs mailing list