Hi,
I am trying to configure orangefs for testing to the following system:
Metadata server /client:
Centos 5 :Linux 2.6.18-194.17.4.el5 #1 SMP Mon Oct 25 15:50:53 EDT 2010 x86_64
x86_64 x86_64 GNU/Linux
Data server:
Debian Lenny :Linux 2.6.17.14 #16 PREEMPT Sat Feb 5 12:45:55 UTC 2011 armv5tejl
GNU/Linux
Data server configuration options:
./configure --prefix=/shares/internal/PUBLIC/orange
Metadata server /client configuration options:
./configure --prefix=/state/partition1/orange
--with-kernel=/usr/src/kernels/2.6.18-194.17.4.el5-x86_64/
The compilation and installation was successful.
"bin/pvfs2-ping -m /mnt/orange" was also successful.
But every time i try to "pvfs2-cp" a file i am receiving a
"trove_write_callback_fn: I/O error occurred" at the data server side(the same
happens with pvfs-2.8.2)
I tried to reconfigure orangefs with many different options but nothing seams
to works.
I also tried "alt-aio" "null-aio" "directio" but nothing happened.
It seams that something strange happens with the data server(ARM processor -
200MHZ,32MB RAM)because both orangefs and pvfs2 works fine when i am using as
data servers intel or AMD processors.
Do you have any idea what is wrong ?
i really need to install orangefs at this ARM processor.
Thanks in advance
Christos
ps: i send you some log files from the data server
alt-aio
--------
PVFS2 Server version 2.8.4-orangefs starting.
[E 07/12/2011 07:38:30] trove_write_callback_fn: I/O error occurred
[E 07/12/2011 07:38:30] handle_io_error: flow proto error cleanup started on
0x154a88: No such file or directory
[E 07/12/2011 07:38:30] handle_io_error: flow proto 0x154a88 canceled 0
operations, will clean up.
[E 07/12/2011 07:38:30] handle_io_error: flow proto 0x154a88 error cleanup
finished: No such file or directory
directio
-----------
PVFS2 Server version 2.8.4-orangefs starting.
[E 07/12/2011 07:56:39] dbpf_bstream_direct_write_op_svc: failed to get dspace
attr for bstream: (error=-1073742082)
[E 07/12/2011 07:57:09] job_time_mgr_expire: job time out: cancelling flow
operation, job_id: 505.
[E 07/12/2011 07:57:09] fp_multiqueue_cancel: flow proto cancel called on
0x113f00
[E 07/12/2011 07:57:09] fp_multiqueue_cancel: I/O error occurred
[E 07/12/2011 07:57:09] handle_io_error: flow proto error cleanup started on
0x113f00: Operation cancelled (possibly due to timeout)
bin/pvfs2-ping -m /mnt/orange/
(1) Parsing tab file...
(2) Initializing system interface...
(3) Initializing each file system found in tab file: /etc/fstab...
PVFS2 servers: tcp://nas-0-1:3334
Storage name: pvfs2-fs
Local mount point: /mnt/orange
/mnt/orange: Ok
(4) Searching for /mnt/orange/ in pvfstab...
PVFS2 servers: tcp://nas-0-1:3334
Storage name: pvfs2-fs
Local mount point: /mnt/orange
meta servers:
tcp://ikaros:3334
data servers:
tcp://nas-0-1:3334
(5) Verifying that all servers are responding...
meta servers:
tcp://ikaros:3334 Ok
data servers:
tcp://nas-0-1:3334 Ok
(6) Verifying that fsid 883136527 is acceptable to all servers...
Ok; all servers understand fs_id 883136527
(7) Verifying that root handle is owned by one server...
Root handle: 1048576
Ok; root handle is owned by exactly one server.
=============================================================
The PVFS2 filesystem at /mnt/orange/ appears to be correctly configured.
--
Christos Filippidis
Institute of Nuclear Physics
National Center for Scientific Research "Demokritos"
Patr. Grigoriou & Neapoleos Str
153 10 Agia Paraskevi Attikis
Athens, Greece
Tel:+30 2106503425
http://www.inp.demokritos.gr/~filippidisx/
_______________________________________________
Pvfs2-users mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-users
