1 / 34

INT408: Migrating from TDM to e-Biz Impact

INT408: Migrating from TDM to e-Biz Impact. Irene Sharpshair Manager Sharpsim@healthall.com Alex Bear Programmer August 15-19, 2004. Introduction. Irene Sharpshair Manager of Databases, Integrated Services & Web The Health Alliance of Greater Cincinnati 3200 Burnet Avenue

deanne
Télécharger la présentation

INT408: Migrating from TDM to e-Biz Impact

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. INT408: Migrating from TDM to e-Biz Impact Irene Sharpshair Manager Sharpsim@healthall.com Alex Bear Programmer August 15-19, 2004

  2. Introduction Irene Sharpshair Manager of Databases, Integrated Services & Web The Health Alliance of Greater Cincinnati 3200 Burnet Avenue Cincinnati, Ohio 45229 Sharpsim@healthall.com Alex Bear Programmer The Health Alliance of Greater Cincinnati 3200 Burnet Avenue Cincinnati, Ohio 45229

  3. Agenda • About The Health Alliance • About the interface engine • Upgrade Process (TDM to Impact 5.3.3.3) • Upgrading to 5.4.1 • Summary • Q & A

  4. About The Health Alliance General Information • The Health Alliance covers the entire Greater Cincinnati area including northern Kentucky • Six hospital facilities including teaching and acute care • Only Level 1 trauma center in area • Air care services • Only heart transplantation program • Surgery Centers • Physician Group

  5. About The Health Alliance - Statistics • No. of Employees: 13,288 • No. of Beds: 1,562* • Admissions (excl. Newborns): 90,414* • Births: 8,423* • Outpatient Registrations: 826,698 (excl. Emergency)* • Emergency Room Visits: 270,179* * July 1, 2002 through June 30, 2003

  6. About The Interface Engine • Started with CAI in 1993 • No. of Interfaces: 117 (as of 6/25/04) • No of transactions processed (March 2004): 14,801,521 • No batch interfaces • Interface Engine for outside Lab referral organization • I.E. sends Lab results to Doctors’ offices • I.E. sends ADT’s, Lab and Radiology results to local CHN (Community Healthcare Network); transcriptions to follow • Three full-time programmers • No charge back (except for outside Lab)

  7. About The Interface Engine - Hardware • 2 L2000 HP-UX 11 servers (multi processor) • 5 Gigabytes of RAM per server • XP1024 disk array used for fail over/disaster recovery • Primary box located at Alliance Business Center • Fail-over box located off site at one of the hospitals • Separate HP box for testing • Kernel Parameters • maxdsiz 483386 • maxssiz 20480 • maxtsiz 16384

  8. Upgrade Process - Preparation Steps • Reviewed documented differences • Staffing and Training • Hardware Requirements • Develop Cluster Design & Implementation Process • Develop Timeline • Document changes for Fail Over/Disaster Recovery • Train Operations

  9. Upgrade Preparation- Review Differences • Reviewed documented differences • No TDM changes • Protocol Object • “protomain” automatically started in 4.x • Must use “.start()” method in 5.x in clinit() function • File objects • No longer supported in DFC calls • clfile argument must be removed from the route_recx and servproc calls • Communication Objects • Must shutdown communications object with method call (comm.kill() or comm.die()) • comm.send(&”KILL”) and comm.send(&”DIE”) are no longer supported

  10. Upgrade Preparation – Review Differences • Reviewed documented differences cont’d • Servayt • Change suspendDFC( ) to clSuspend( ) • Change releaseDFC( ) to clRelease( ) • Serial number has been changed to string for ease of use • FORMAT – 12345:Prodname • Old “Explode/Combine” method of dfc is no longer supported • Servayt argument list should now be string • int [in,flavor] AYTflav, string [out] *serial, int [in] sfmFlav

  11. Upgrade Preparation – Review Differences • Reviewed documented differences cont’d • Servproc • Change suspendDFC( ) to clSuspend( ) • Change releaseDFC( ) to clRelease( ) • Remove the clfile object from the servproc entry point • Change the *ErrText parameter from [in,out] to just [out] • Serial number has been changed to string for ease of use • FORMAT – 12345:Prodname • Old “Explode/Combine” method of dfc is no longer supported • servproc argument list should now be string • int [in,flavor] Flav, string [in,out] *serial, char [in] src, blob [in] DataBuf, string [in] tranID, string [out] *errText, string [in] fKey, int [in] opts, char [in] priority, int [in] sfmFlav

  12. Upgrade Preparation – Review Differences 1) Reviewed documented differences cont’d • C-Based aims issues • C-Based command-line client issues • ODL command-line clients • Clfile objects from SFM are no longer available due to removal from route_recx

  13. UpgradePreparation - continued 2)Staffing and Training • Consulting firm hired for assistance in training/project 3)Hardware Requirements • Sybase concluded that current hardware would meet our needs • Underestimated memory needs – added 2 gig for a total of 5 • Sybase has developed a more accurate method to determine needs • Develop Cluster Design & Implementation Process • Decided to implement a cluster at a time • Added a delivery aim from currentproduction environment to new eBIZ cluster

  14. Original TDM Production Environment

  15. Original Interface Design

  16. UpgradePreparation - continued 5) Develop Timeline Original Implementation Timeline • Process started end of June • First clusters live on 8/12/03 • All clusters to be live on 9/30/03

  17. 6/23/03 Installed V 5.3.1 & Patch 5.3.1.8 MAXDSIZE error 7/3/03 Third cluster installed on test box Could not monitor through global console with standard SNMP ports DFCSRV functions for inter-cluster communications did not work Increase parm to 483386 Moved from NT to Windows PC Used Acquisition & Delivery aims Actual Implementation Process

  18. 7/27/03 Installed V 5.3.2.6 CNC commands used to start the clusters were changed DFCSRV function now worked 8/12/03 FIRST CLUSTER LIVE Sybase advised us on new commands syntax (ims wrapper script now used) Replaced Acquisition & Delivery aims between clusters with DFCSRV Actual Implementation Process

  19. 8/20/03 Controller failed when it went into re-try mode. Unable to pause a destination while in re-try 8/28/03 Controller problems continued Sybase modified our aims which were given to us from consultant clRelease added to open flow Restart logic was moved from close flow to servproc clQuit removed from No Connect and clSetAimStatus added Sybase fixed in 5.3.3.3 and we installed on 9/20 Discovered 64-bit kernel parameters had been changed and not the 32-bit. maxdsiz, maxssiz, maxtsiz Actual Implementation Process

  20. 9/1/03 – 9/30/03 Upgrade to V 5.3.2.8 Implemented 3 more clusters Upgraded to V 5.3.3.3 Issues starting multiple SNMPs Controller problems 10/2/03 Problems with one high volume cluster (We had combined 2 high-volume SFM’s in CAI to 1 cluster and 1 controller) Nov-Dec Received new SNMP binary 10/05 Implemented pager alerts and auto-restarts Split single cluster into two clusters Brought up 7 more clusters Memory issues started Actual Implementation Process

  21. 12/4/03 Discovered Lab was receiving Orders before ADT’s (It had 1 destination and 3 production objects) 1/4/04 Memory usage was at 87% and we had 3 very large clusters to go 2/28/04 Cluster design changed Added Security functionality Switched from parallel processing to serial processing Put Lab in separate SFM so other 5 destinations not impacted Project put on hold while more memory was ordered Added multiple controllers to clusters (Acq, Del, SFM) Design change gave us ability to do security at the level we wanted for Operations Actual Implementation Process

  22. 4/8/04 Memory added (2 gig) for a total of 5 gig All remaining clusters implemented Actual Implementation Process

  23. Upgrade Preparation - continued 6) Document changes for Fail Over/Disaster Recovery • New scripts were written to start Impact 5.3 and added to package startup • Script starts SNMP daemons • Verifies select daemons are running • Calls clusters startup script • Run level of package changed from RC2.d to RC4.d • Directory permissions changed • /tmp/.AgentSockets 7) Train Operations • Exception monitoring • New manual • Training sessions

  24. 4/19/04 Started work on 5.4. May-June Directories changed and Wrapper Script names changed DFCSRV properties not viewable in version 5.3 with 5.4 installed on PC SNMP daemons not able to resolve a response timeout issue (recurrence of 5.3 issue) 6/28/04 Applied to test server to evaluate Made changes to custom scripts to allow for changes Implementation plan changed to migrate clusters with DFCSRVs first Replace “snmpdm” binary with previous binary received from Sybase R&D GO LIVE ON 5.4.1 – all systems Upgrading to 5.4.1

  25. Original TDM Production Environment

  26. Original Interface Design

  27. Final Interface Design 5.3.3.3

  28. Final Interface Design 5.4.1

  29. SUMMARY • Best Practices • Production Implementation – a cluster at a time with 2-3 days in between • Adding more controllers • Adding pager alerts • Adding auto-restart to acquisition aim • Adding Security • Using endpoint simulator • Using multiple global consoles on same PC • Single Delivery and Acquisition aims • Parameters File • Acquisition includes • Source, Debug, sfmflavor, tran, prod, log, restart, acktype, veng • Delivery includes • Debug, Log, SendOk, AERefuse, AckWait, ARRefuse, GCStatus

  30. SUMMARY cont’d • What we kept • Desktop (for Operations) • Automatic ping system • Ancillary owner that should be called • Our ALERT program – used to automate pager alerts • Our MSGLOGS – Contains actual message sent, the ACK received and timestamps • Our program to resend batches of data • What we removed • Custom program to restart systems now done in Global Console • Custom program which displayed all systems on one screen replaced with Global Console • Custom program to check if this is a known downtime for the system. If not, restart the system. This is now done through the Configurator.

  31. SUMMARY cont’d • Helpful commands • Determines PID of a SNMP daemon: lsof –i:xxxx (xxxx = port) • Performance monitor for HP-UX: glance - Displays top processes on system: top - Prints the route that IP packets take traceroute xxx.xx.xxx (IP address) • Script to track memory usage by watchdog shell (from Sybase) process • Our Scripts • Alerts script • Pending Check script • Engine paging

  32. Open Issues • Case 11020297* • Memory leak in ODL • Currently an issue with Sybase R&D CR #352451 • Case 11023123* • Memory leak in CNC • Currently an issue with Sybase R&D CR #245796 • Case 11047078* • Global console shows destination as active when it is paused • Currently an issue with Sybase R&D CR #359559 *These issues exist in V5.3 & V5.4.1 and as of 6/25/04 the cases were still open with Sybase

  33. Questions ????

More Related