Post date: Apr 16, 2012 2:33:59 PM
28 April
All the Padova machines are down for an electronic problem in the computing room...
27 April
Condor (version 7.6.6 -- 2012) on the machines
Now:
- photon00 condor manager
- photon01->09 + 12 + 13 +VM01 CPU --> total 82 cores (photon02 and 03 missing since they are switched off)
26 April
New server cores is on:
photon13 -> new server 32 core
photonvm01 -> virtual machine
24 April
New server disks with /home, /soft and data18/19/20 is on and mounted on all the machines.
18 April
UPDATE:
unfortunately the architecture that we have agreed two days ago with Fabio cannot be applied.
Therefore, we have decided this second solution:
- 4 Tb for home and soft in raid 10 (= super-secure)
- 6 Tb for the three new disks in raid 5 (secure as our other disks)
----------
Yesterday the new server has been installed on the rack (with some problems due to the fact that our rack is a bit old). Therefore, today Fabio will proceed to the mounting.
Some additional information and thoughts/suggestions:
- the old disk system is maintained as it is.
- we suggest to use /home to store general personal stuff plus data for the current analysis (super-backupped) and then move old analysis elsewhere
- we plan to have a central repository for MC for the whole group to avoid that each one download and reconstruct all the MC (afterward one can play with symoblic links)
- the server name will likely be photon13, and will not be accessible for running things.
16 April:
We are installing the new server (Villi, Elisa, Fabio B.).
Some info: the server has 8 disks, 2 Tb each + two small disks for the system.
Proposal:
- 1 disk (2 Tb) with /home and /soft + 1 disk for mirror (veeery secure!);
- the other 5 disks are: 4 disks for data + 1 disk for raid 5 (8 new fresh Tb for our analyses!);
- finally 1 disk "hot spare"-> if a disk breaks, this will replace the missing disk;
Additional info
- this server will not be used for running jobs but only as storage;
- we will install a 64 bit linux version in it to optimize its usage.
Moreover (Cornelia):
This morning we realised that there were some problems accessing the photon machines. Connections to the machines establihsed before this morning were still active and usuable. We contacted the support in order to find out what was going on. They informed us after some investigations that one of the 3 servers taking care of the user and password authentication was down, in fact exactly the one we are using for accessing our machines. After some hours the server was up again and the problem thus solved.