View source for RAL-LCG2-Tier-1

= RAL-LCG2-Tier-1 =
== Topic: HEPSPEC06 ==
<br />
<br />
{| border="1"
|+ RAL Tier-1
|-
|Generation||config, CPU||OS||Kernel||32/64||mem||gcc|| Total  || Per Core || Notes
|-
|-style="background:green;color:black"
|2007 Streamline||dual Intel E5410 @ 2.33GHz||SL5.4|| - ||32bit on 64bit OS||16GB||gcc version||66.537|| 8.317
|-
|-style="background:green;color:black"
|2007 Clustervision||dual Intel E5440 @ 2.83GHz||SL5.4|| - ||32bit on 64bit OS||16GB||gcc version||75.617|| 9.452
|-
|-style="background:green;color:black"
|2008 Streamline||dual Intel L5420 @ 2.50GHz||SL5.4|| - ||32bit on 64bit OS||16GB||gcc version||69.547|| 8.693
|-
|-style="background:green;color:black"
|2008 Viglen||dual Intel E5420 @ 2.50GHz||SL5.4|| - ||32bit on 64bit OS||16GB||gcc version||70.760|| 8.845
|-
|-style="background:green;color:black"
|2009 Viglen||dual Intel E5520 @ 2.26GHz||SL5.4|| - ||32bit on 64bit OS||24GB||gcc version||92.593||11.574|| Average All Units
|-
|-style="background:green;color:black"
|2009 Streamline||dual Intel E5520 @ 2.26GHz||SL5.4|| - ||32bit on 64bit OS||24GB||gcc version||92.170|| 11.521|| Average All Units
|-
|-style="background:green;color:black"
|2010 Clustervision/Dell||dual Intel X5650 @ 2.66GHz||SL5.4|| - ||32bit on 64bit OS||48GB||gcc version||166.410|| 13.868|| Average All Units
|-
|-style="background:green;color:black"
|2010 Viglen||dual Intel X5650 @ 2.66GHz||SL5.4|| - ||32bit on 64bit OS||48GB||gcc version||156.030|| 13.003|| Average All Units
|-
|-style="background:green;color:black"
|2011 Viglen||dual Intel E5645 @ 2.40GHz||SL5.4|| - ||32bit on 64bit OS||48GB||gcc version|| - || - || Average All Units
|-
|-style="background:green;color:black"
|2011 Dell||dual X5650 @ 2.66GHz||SL5.4|| - ||32bit on 64bit OS||48GB||gcc version|| - || -|| Average All Units
|-
|}
<br />
<br />
{{KeyDocs|responsible=Rob Harper|reviewdate=2012-03-15|accuratedate=(never)|percentage=0}}
<br />
<br />
== Topic: Middleware_transition ==
<br />
====gLite3.1/SL4====
<br />
WMS  - 3 nodes v3.1.32-0.slc4<br />
<br />
LCG-CE - 1 node v3.1.37-0 (planned to be migrated to glite3.2 CREAM, not scheduled yet)<br />
<br />
PROXY - v3.1.29-0<br />
<br />
====gLite3.2/EMI====
<br />
APEL - v3.2.5-0.sl5<br />
<br />
ARGUS - n/a<br />
<br />
BDII_site - 3 nodes v3.2.10-1.sl5<br />
<br />
BDII_top - 5 nodes v3.2.10-3.sl5<br />
<br />
CE (CREAM/LCG) - 4 CREAM nodes v3.2.10-0.sl5, 1 CREAM node v3.2.6-0.sl5<br />
<br />
FTS - v3.2.1-2.sl5 (Oracle backend)<br />
<br />
glexec - v3.2.2-2.sl5<br />
<br />
LB - 2 nodes v3.2.12-11.sl5<br />
<br />
LFC - 7 nodes v3.2.7-2 (Oracle backend)<br />
<br />
SE - Castor 2.1.10-0<br />
<br />
UI - v3.2.10-1.sl5<br />
<br />
VOBOX - v3.2.11-0.sl5<br />
<br />
WMS - n/a<br />
<br />
WN - v3.2.7-0<br />
<br />
====Comments====
<br />
Plan to virtualize further Grid Services deployments using Microsoft Hyper-V<br />
<br />
Regarding migration to EMI/UMD, no work yet, but would prefer to have access to some installation and maintenance recipes supported by developers (similar to what does exist for gLite).<br />
<br />
== Topic: Protected_Site_networking ==
<br />
*Tier1 is a subnet of the RAL /16 network<br />
*Two overlaid subnets: 130.246.176.0/21 and 130.246.216/21<br />
*Third overlaid /22 subnet for Facilities Data Service<br />
*To be physically split later as traffic increases<br />
*Monitoring: Cacti with weathermaps <br />
*Site SJ5 link: 20Gb/s + 20Gb/s failover direct to SJ5 core two routes (Reading, London)<br />
*	T1 <-> OPN link: 10Gb/s + 10Gb/s failover, two routes •	T1 <-> Core 10GbE •	T1 <-> SJ5 bypass: 10Gb/s •	T1 <-> PPD-T2: 10GbE<br />
*	Limited by line speeds and who else needs the bandwidth<br />
<br />
[[Image:tier1-network.jpg]]<br />
<br />
<br />
== Topic: Resiliency_and_Disaster_Planning ==
<br />
<br />



*** This section intentionally left blank 


== Topic: SL4_Survey_August_2011 ==
<br />
<br />
<br />



*** This section intentionally left blank 


== Topic: Site_information ==
<br />
[[Memory]] <br />
<br />
1. Real physical memory per job slot:<br />
<br />
All WNs have 2GB/core (1 job slot per core).<br />
<br />
2. Real memory limit beyond which a job is killed:<br />
<br />
Dependent on queue : 500M,700M,1000M,2000M,3000M<br />
<br />
3. Virtual memory limit beyond which a job is killed:<br />
<br />
no limit<br />
<br />
4. Number of cores per WN:<br />
<br />
4 or 8 depending on hardware.<br />
<br />
Comments:<br />
<br />
[[Network]]<br />
<br />
1. WAN problems experienced in the last year:<br />
<br />
2. Problems/issues seen with site networking:<br />
<br />
3. Forward look:<br />
<br />
Plan for doubled 10GbE (20Gb/s) for internal links and doubling of existing links as needed.<br />
<br />
Comments:<br />
<br />
== Topic: Site_status_and_plans ==
<br />
<br />
[[SL5 WNs]] <br />
<br />
Current status (date): 19/11/2009 All LHC accessible WNs are SL5, <br />
<br />
Planned upgrade: None<br />
<br />
Comments: <br />
<br />
[[SRM]]<br />
<br />
Current status (date): <br />
<br />
Planned upgrade:<br />
<br />
Comments:<br />
<br />
[[SCAS/glexec]]<br />
<br />
Current status (date): 10/05/2011 SCAS deployed, intend to look at ARGUS but delayed due to staff changes.<br />
<br />
Planned deployment:<br />
<br />
Comments:<br />
<br />
[[CREAM CE]]<br />
<br />
Current status (date): 10/05/2011 CREAM CEs available for all VOs, planning reinstallation of remaining 2 lcg-CEs<br />
<br />
Planned deployment:<br />
<br />
Comments:<br />
<br />
<br />