Page History

• Migration to Java 17
  • Followup Jan 19
  • Learned OS and Tomcat due for upgrade as well
    • We want no responsibility for this - we are agnostic
    • Too many managers and not enough techs involved
  • Java 17 installed on Dev
    • jconsole not fully supported by Corretta
  • Still need Tomcat updated
    • Can we update ejb ourselves

Kim's Proposal

• Programmatic options
• CRM to filter messages
  • Would be a service that sets up a queue of real issues
  • "High noise, low signal"
  • Cost? $5-30/user.  What is a user?
• Inventory public facing pages for email address
  • Can we get rid of Contact Us page?
  • Send email to user to confirm before sending onwards to final target?
• Source appears to be the Contact Us page, as it is being sent to both NCIThesaurus and NCICBIIT
  • A lot is coming from an .ru address
  • We have a captcha on that page

• Still multiple issues on Prod
  • Scott creating Jira issues as items identified
  • Possibly check email archives for previous logs
  • Latest tag is on QA and no show-stoppers seen
• J console up and running on Dev
  • Additional libraries needed to be added to Corretta
  • Testing can start now - stress testing using Selenium
  • Still trying to replicate the messages coming in from Prod
  • Possibly people walking the hierarchy - Hierarchy pages seeing lots of traffic

• Initial "Conversions".
• DataQA after Prod
  • database on machine - migrate first?
  • Test performance of old loaders vs new database.
  • Other apps follow DB - determine if needs to be co-located
  • Does script need to be updated?

ehcache issues on Meta, similar to what we have seen previously on NCI Terms 

Question: is there a debugging or logging ticket for PROD? 

• There is an open question of where the logging would be turned off.
• Scott believes it is a flag somewhere. He has tried tons of approaches but has not been able to figure out where controlled. Scott will write ticket for debugging.
• Scott tried Kim's approach, suggestions but was not fruitful.

Both tags on QA have passed. Will update once deployment happens.

Jason put in ticket to have `lbpatch.jar` on STAGE, but Systems could not find that jar.

TODO: file a ticket to look for the lbpatch.jar  on PROD.

•  Scott to examine

EVS Prod is down (during mtg).

• maybe Apache is being restarted?
• Ticket: restart application (Eric Gavua), "Yes, it did go down. I have restarted the service"
• request to get htop and du -hs { } metrics on PROD
• p805 log, there are 2 restarts, 9:19am, and back to working is around 10:10am. (the 40 min is due to caching)
• no caching info are seen in logs; no evidence that it is coming back up (caching) seen in logs.
• it is failing immediately.
• The "maintenance" page appears with the full url, but hostname doesn't show this...
• It may be an issue of lack of server resources (CPU, Memory, disk space, swap); the index building on other server is much less.

Can add and remove users through Service Now.