Hi Maarten,
Checked also this one.
GLOBUS variables mentioned are not set on WNs. The same problem happened
with local dpm - and there is no
firewall in between - might be other problem
timeouts are set what means that lcg-cp shoud respect them when calling
copyfilex(), seems like an error
Here's how it looks for biomed:
ps:
47007 30069 22801 0 18:22 ? 00:00:00 lcg-cp --vo biomed
--checksum --connect-timeout 300 --sendreceive-timeout 300
--bdii-timeout 300 --srm-timeout 300
srm://se03.esc.qmul.ac.uk:8444/biomed/a2d6fafd-eeee-4d18-b4af-24d88385c8a1/persistent/493a53ad052e59e4615207cd239fc1184376d932_492278f9-7634-4afe-9bc9-ac70b7f66363.rep
file:/scratch/23292999.batch.grid.cyf-kr.edu.pl/CREAM522320835/ws15690/openmole.tar.gz.gz
[root@n16-4-32 ~]# stat
/scratch/23292999.batch.grid.cyf-kr.edu.pl/CREAM522320835/ws15690/openmole.tar.gz.gz
File:
`/scratch/23292999.batch.grid.cyf-kr.edu.pl/CREAM522320835/ws15690/openmole.tar.gz.gz'
Size: 0 Blocks: 0 IO Block: 4096 regular
empty file
Device: 802h/2050d Inode: 10584390 Links: 1
Access: (0644/-rw-r--r--) Uid: (47007/biomed007) Gid: (47000/ biomed)
Access: 2012-09-12 18:23:21.000000000 +0200
Modify: 2012-09-12 18:23:21.000000000 +0200
Change: 2012-09-12 18:23:21.000000000 +0200
[root@n16-4-32 ~]# pstack 30069
#0 0x00000038f70cc223 in __select_nocancel () from /lib64/libc.so.6
#1 0x00002ab317baf5d4 in globus_l_xio_system_poll () from
/opt/globus/lib/libglobus_xio_gcc64dbg.so.0
#2 0x00002ab3198b0534 in globus_callback_space_poll () from
/opt/globus/lib/libglobus_common_gcc64dbg.so.0
#3 0x00002ab316b7e0c2 in copyfilex () from /opt/lcg/lib64/liblcg_util.so.1
#4 0x00002ab316b76a8a in lcg_cp5 () from /opt/lcg/lib64/liblcg_util.so.1
#5 0x0000000000401a58 in main ()
[root@n16-4-32 ~]# lsof -p 30069 | grep -e TCP
lcg-cp 30069 biomed007 7u IPv4 152800688 TCP
n16-4-32.local:53426->se04.esc.qmul.ac.uk:gsiftp (ESTABLISHED)
On 12.09.2012 21:14, Maarten Litmaath wrote:
> Hi Lukasz,
>
>> Unfortunately clients are still hanging even after coresponding server
>> has been killed.
>
> Check their open connections with "lsof"? Firewall/router issue?
>
> Maybe this old problem:
>
> https://wiki.egi.eu/wiki/Tools/Manuals/TS23
|