Xrootd Monitoring and Control Harsh Arora CERN

Setting Up Service  Monalisa Service  Monalisa Repository  Test Xrootd Server  ApMon Module

Collecting Xrootd Parameters Monalisa Service External Storage Xrootd System/ Network Info Sends monitoring data Control Xrootd Services

Proposed Architecture to Monitoring & Control the Xrootd in ALICE Long History DB MonALISA Network Monitoring ApMon XROOTD SERVER System Monitoring MonaLisa Xrootd Repository Aggregated Data Alerts Actions Iosif Legrand June 2012 Control Module Control / Update MonALISA Network Monitoring ApMon XROOTD SERVER System Monitoring Control Module MonALISA Network Monitoring ApMon XROOTD SERVER System Monitoring Control Module

  Run a MonALISA service on each host running an XROOTD server ( a new dedicated group )   Control the xrootd server   Start / Stop / Update   Dynamically Change conf parameters   Collect Monitoring information from the xrootd servers using the new & improved monitoring functionality from xrootd   Perform full network measurements and tests (RTT, available bandwidth, topology )   Monitor the storage system used by xrootd servers   Create a dedicated MonALISA repository 5 Iosif Legrand June 2012 Proposed Architecture to Monitoring & Control the Xrootd in ALICE

Active Available Bandwidth measurements between all the ALICE grid sites (2) Iosif Legrand October Iosif Legrand June 2012 ALICE measurements between all VO Boxes Important to have similar Data for Xrootd servers

7 Advantages of Using a dedicated MonALISA group to monitor and control all the XROOTD servers  Easy to maintain and update a critical service for Offline computing  Significantly improve the monitoring information and will help to better understand the way storage system are used in the ALICE Computing model (Distributed Storage for Data)  Control the xrootd servers and can dynamically configure systems based on how they are used.  Monitor the true network connectivity  Monitor the real storage used by xrootd servers  Monitor the connection and transfer per client / job  Can really help in developing optimization procedures for complex data flows Iosif Legrand June 2012

% of CPU Used by Monalisa

Parameters Collection and Verification

% of time spent by CPU in User Mode Monalisa Host Monitoring Reported by Xrootd

Parameters Open_files Srv_rd_mbytes Srv_wr_files

Parameters(Cont.) Srv_rd_filesSrv_wr_mbytes Space_free

Average Load over last 15 min. Load15

Ambiguous Parameters  Connected Clients  Number of Jobs Requiring a Thread

Dedicated MLRepository

Detailed Monitoring per Client  Monitoring per Client  Patch to xrootd-3.1.0

Client Monitoring Example Number of Bytes

Additional Information provided by monitoring each client from the xrootd server  Global Map of Data Transfers  Debug individual clients  Understand better the effect of long RTT

Preparing Correlations Plots  Bad xrootd Server Detection  High CPU Usr with Low Disk Throughput  -ve correlation b/w Traffic and Disk IOUtil(%)  High Disk IOUtil(%) with Less Read Write Rate – Probably Damaged Disk  High Load Value with Less number of clients

Monalisa on Each Server  Local CPU Monitoring Info  Understanding Network Topology  Traffic Information  Monitoring directly the storage system  Control

Future Plan

Control via Monalisa  Start/Stop xrootd server  Upgrade

Problems  Latest version of xrootd  Incompatible with xrootd-ftsofs-1.1.0

Thank You