[Pvfs2-users] PVFS2 Installation

Jeff Pummill jpummil at uark.edu
Tue Apr 18 07:42:39 EDT 2006


And things continue to improve!

[jpummil at prospero bin]$ ./pvfs2-ping -m /mnt/pvfs2

(1) Parsing tab file...

(2) Initializing system interface...

(3) Initializing each file system found in tab file: /etc/pvfs2tab...

   /mnt/pvfs2: Ok

(4) Searching for /mnt/pvfs2 in pvfstab...

   PVFS2 servers: tcp://comp-pvfs-0-24.local:3334
   Storage name: pvfs2-fs
   Local mount point: /mnt/pvfs2

   meta servers:
   tcp://pvfs2-meta-server-0-0:3334

   data servers:
   tcp://comp-pvfs-0-24:3334
   tcp://comp-pvfs-0-25:3334
   tcp://comp-pvfs-0-26:3334
   tcp://comp-pvfs-0-27:3334
   tcp://comp-pvfs-0-28:3334
   tcp://comp-pvfs-0-29:3334

(5) Verifying that all servers are responding...

   meta servers:
   tcp://pvfs2-meta-server-0-0:3334 Ok

   data servers:
   tcp://comp-pvfs-0-24:3334 Ok
   tcp://comp-pvfs-0-25:3334 Ok
   tcp://comp-pvfs-0-26:3334 Ok
   tcp://comp-pvfs-0-27:3334 Ok
   tcp://comp-pvfs-0-28:3334 Ok
   tcp://comp-pvfs-0-29:3334 Ok

(6) Verifying that fsid 1594492348 is acceptable to all servers...

   Ok; all servers understand fs_id 1594492348

(7) Verifying that root handle is owned by one server...

   Root handle: 1048576
Warning: non PVFS2 error code (16):
PVFS_mgmt_setparam_all: Invalid argument
Failure: failed to check root handle.
[jpummil at prospero bin]$ ./pvfs2-statfs -m /mnt/pvfs2

aggregate statistics:
---------------------------------------

        fs_id: 1594492348
        total number of servers (meta and I/O): 7
        handles available (meta and I/O):       3681400536
        handles total (meta and I/O):           3681400536
        bytes available:                        25643876352
        bytes total:                            35269730304

NOTE: The aggregate total and available statistics are calculated based
on an algorithm that assumes data will be distributed evenly; thus
the free space is equal to the smallest I/O server capacity
multiplied by the number of I/O servers.  If this number seems
unusually small, then check the individual server statistics below
to look for problematic servers.

meta server statistics:
---------------------------------------


I/O server statistics:
---------------------------------------

server: tcp://comp-pvfs-0-24:3334
        RAM bytes total  : 1059573760
        RAM bytes free   : 312512512
        uptime (seconds) : 509398
        load averages    : 66624 66560 65536
        handles available: 613566756
        handles total    : 613566756
        bytes available  : 4274118656
        bytes total      : 5878288384
        mode: serving only I/O data

server: tcp://comp-pvfs-0-25:3334
        RAM bytes total  : 1059573760
        RAM bytes free   : 317526016
        uptime (seconds) : 509296
        load averages    : 65536 66016 65536
        handles available: 613566756
        handles total    : 613566756
        bytes available  : 4274139136
        bytes total      : 5878288384
        mode: serving only I/O data

server: tcp://comp-pvfs-0-26:3334
        RAM bytes total  : 1059573760
        RAM bytes free   : 345608192
        uptime (seconds) : 509243
        load averages    : 67680 67008 65536
        handles available: 613566756
        handles total    : 613566756
        bytes available  : 4274171904
        bytes total      : 5878288384
        mode: serving only I/O data

server: tcp://comp-pvfs-0-27:3334
        RAM bytes total  : 1059573760
        RAM bytes free   : 315330560
        uptime (seconds) : 509187
        load averages    : 69920 70304 66432
        handles available: 613566756
        handles total    : 613566756
        bytes available  : 4274118656
        bytes total      : 5878288384
        mode: serving only I/O data

server: tcp://comp-pvfs-0-28:3334
        RAM bytes total  : 1059573760
        RAM bytes free   : 321171456
        uptime (seconds) : 509141
        load averages    : 68128 69760 66144
        handles available: 613566756
        handles total    : 613566756
        bytes available  : 4274126848
        bytes total      : 5878288384
        mode: serving only I/O data

server: tcp://comp-pvfs-0-29:3334
        RAM bytes total  : 1059573760
        RAM bytes free   : 296800256
        uptime (seconds) : 509101
        load averages    : 65536 66624 65536
        handles available: 613566756
        handles total    : 613566756
        bytes available  : 4273979392
        bytes total      : 5878288384
        mode: serving only I/O data

Next, I tried to do a pvfs2-cp. This seemed to have issues....

[root at prospero bin]# ./pvfs2-cp 
/home/jpummil/libgfortran-4.0.2-8.fc4.i386.rpm 
/mnt/pvfs2/libgfortran-4.0.2-8.fc4.i386.rpm
[11:41:50.473719] Object Type mismatch error: Not a directory
[11:41:50.473971] PINT_sm_common_parent_getattr_failed: Not a directory
PVFS_sys_create: Not a directory
Could not open /mnt/pvfs2/libgfortran-4.0.2-8.fc4.i386.rpm

[jpummil at prospero bin]$ ps -ef | grep pvfs
root      6815     1  0 Apr12 ?        00:00:00 /usr/sbin/automount 
--timeout=60 --ghost /mnt/pvfs2 file /etc/auto.pvfs2
root      6869     1  0 Apr12 ?        00:00:00 
/opt/pvfs2/sbin/pvfs2-client -p /opt/pvfs2/sbin/pvfs2-client-core
root      6870  6869  0 Apr12 ?        00:00:01 pvfs2-client-core -a 5
jpummil  18952 15737  0 11:39 pts/4    00:00:00 grep pvfs


Jeff F. Pummill



Robert Latham wrote:
> On Tue, Apr 18, 2006 at 05:55:43AM -0500, Jeff Pummill wrote:
>   
>> I have re-run pvfs2-genconfig using the /pvfs2-data directory found on 
>> the I/O and meta nodes as the storage target.
>>
>> Oddly, the I/O and meta nodes appear to be running the pvfs2 client-core 
>> rather than the server.
>>     
>
> Thanks for including your steps.  You've got the right idea, but there
> are a few things you are missing before 'pvfs2-ping' will work:
>
> You've got your config files now (/etc/pvfs2-fs.conf and
> seven files like /etc/pvfs2-server.conf-comp-pvfs-0-XX ), so that's a
> good start.   When i say 'comp-pvfs-0-XX' i mean "the name of an I/O
> or metadata server"
>
> You'll have to somehow get those files from prospero to the other
> nodes.
> http://www.pvfs.org/pvfs2/pvfs2-quickstart.html#sec:server-config
> describes one approach using 'scp', but it's probably easiest to just
> "scp /etc/pvfs2-* comp-pvfs-0-XX:/etc/" for each of your servers.
>
> Then on each node, you need to create the storage space with the '-f'
> option to pvfs2-server:
>
>     pvfs2-server -f pvfs2-fs.conf pvfs2-server.conf-comp-pvfs-0-XX
>
> After you create the storage space, then you can start the servers.
> Again, do this on each node (note, there's no -f option):
>
>    pvfs2-server pvfs2-fs.conf pvfs2-server.conf-comp-pvfs-0-XX
>
> Once all the servers are running (one on each of your machines) you'll
> be able to run pvfs2-ping.  Once pvfs2-ping reports everything is OK,
> then you can move on to the kerrnel interface.  
>
> ==rob
>
>   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.beowulf-underground.org/pipermail/pvfs2-users/attachments/20060418/9bfcb4d7/attachment.htm


More information about the Pvfs2-users mailing list