JLab RDB-based Archiver

1
JLab RDB-based Archiver

• Matt Bickley
• (work by Anthony Bavuso and Chris Slominski)

2
What? Another Archiver?

• ScalabilityRequirement to support monitoring of
  1M channels
• ManageabilityRequiring no interaction by
  administrators
• Design issues
• Data latency (10-20 minutes from Czar)
• Independent deadbanding (not using ADEL)
• Data loss

3
Data loss from Czar

• Users saw transients in StripTool that were not
  recorded by Czar
• IOC CPU loading less than 30
• Czar system loading less than 20
• Confirmed data loss with a test setup

Archive
Analyzer
Czar
Test IOC
Database
CA Server
FFFF FFFE FFFC . .
FFFF FFFE FFFD FFFC . .
FFFF FFFE FFFC . .
4
MYA Design
User Commands
Watchdog Thread
Copycat Thread
Main Thread
MySQL Database
EPICS Thread
Control System
5
MYA realtime thread scheduling

• Prioritization, highest to lowest
• Channel Access communication
• Database scribe threads
• Client communication, watchdog, etc.
• No loss of data
• If work queue gt 2M events then database
  insertions are prioritized
• Peak queue length 160,000 events
• Average queue length 5,000 events

6
System Hardware

• Dell PowerEdge 2850
• Dual quad-core CPUs
• 16 GB memory
• 8K in 2006 (4K for the memory)
• EonStore RAID box
• 16 300GB SCSI disks (2 TB of storage)
• RAID 01
• 21K in 2006 (17K for the disks)

7
Performance

• 50,000 channels monitored
• 25,000 updates/second
• 4,000 database insertions/second
• History data retrieval rate of 160,000/second(Get
  one day of a 1 HZ channels data in 0.5 seconds)
• Store at least 1 year of data online

8
Whats Next

• Cut over to MYA for operations
• Sandbox archiver
• Users specify archive requests
• Limited data retention (up to 1 month)
• Limited archive request lifetime
• OR
• Archive everything
• 300,000 channels
• Keep the data for 1 week (?)
• Needs more memory, much less disk space