Hi,
I am trying to run the code on small PROOF cluster of 2 machines each with 48 workers.
I have done it successfully on other 2 machines with less number of cores. However, on the
actual cluster when I run xrootd ‘xrootd.log’ shows that loading of rootd protocol for the file
serving fails. Attaching the xroot.log along with configuration and proof.conf file. xroot_log.txt (9.5 KB) xpd_proof_conf.txt (1.4 KB) proof_conf.txt (3.4 KB)
In my xrootd config file,
all.adminpath is set to be /home/maceuser1/easlib/pool/admin.
/home/maceuser1/easlib is actually a soft link in user-home of each machne
which points at respective storage partition on each machine. (The storage
partitions are named differently on each machine ). Is that the
source of the problem ?
Hi,
I do not know if that is the source of the problem, but what is certain is that the daemon of one of the machines is not able to create what it needs under that path
180703 14:54:58 10261 ofs_Config: Unable to create path for /home/maceuser1/easlib/pool/admin/maceuser1/.ofs/posc.log/maceuser1; not a directory
So, I would check permissions and similar things in there.
but what is certain is that the daemon of one of the machines is not able to create what it needs under that path
I noticed this error. However path “/home/maceuser1/easlib/pool/admin/maceuser1/.ofs/posc.log” does exist.
however posc.log is a file generated by daemon. Why daemon is trying to create directory, there I can’t understand.
I am following instructions at https://root.cern.ch/configuring-cluster-workers-pool-disk-each-worker for my configuration file. Now during test setups I connected 2 normal machines and was using the storage of the machines themselves. It worked.
For my actual cluster 2 servers have storage disks mounted at /data3 and /data4 respectively on 2 machines. I will be using machine 1 as redirector as well. Do I need to make some changes for this configuration ?