Before executing significant operational procedures on production systems
09 September 2019 05:47 AM
Before executing significant operational procedures on production systems, such as
- Production Go Live events;
- Major version Upgrades;
- Adding/removing nodes to a cluster;
- Deploying a new application or an application upgrade;
- Thorough testing of any operational procedures on non production systems.
- Opening a ticket with MarkLogic Technical Support to give them a heads up, along with any useful collateral that would help expedite diagnostics of issues if any occur, such as
- The finalized plan & timeline or schedule of the operational procedure.
- A support dump, taken before the operational procedure, in order to record the configuration of the system ahead of time; This may come in handy if an incident occurs as we may want to know the actual changes that had been made. You can create a MarkLogic Server support dump from our Admin UI by selecting the 'Support' tab; select scope=cluster, detail=status only, destination=browser -> save output to disk. Attach the support dump to the ticket as a file either as an email attachment or uploading through our support portal.
- A few days of error logs from before the operational event so that we can determine whether artifacts in the error logs are new or whether they existed prior to the event.
- You can alternatively turn Telemetry on before the event and force an upload of the support dump & error logs.
- Any architecture or design details of the system that you are able to share.
- Please make sure that all individuals who are responsible for the event and who may need to contact the MarkLogic Technical Support team are registered MarkLogic Support contacts. They can register for an account per instructions available at https://help.marklogic.com/marklogic/AccountRequest. They will want to register before the event as ONLY registered support contacts can create a ticket with MarkLogic Technical Support. We do not want registration and entitlement verification to get in the way of the ability to work on an urgent production issue.
- Review the MarkLogic Support Handbook - http://www.marklogic.com/files/Mark_Logic_Support_Handbook.pdf. The following sections in the "HOW TO RECEIVE SUPPORT SERVICES" chapter of the handbook are useful to be acquainted with before an incident occurs
- Section: What to do Prior to Logging a Service Request
- Section: Working with Support
- Section: Escallation Process
- Section: Understanding Case Priority and Response Time Targets
- For urgent issues (production outages), remember that you can raise an urgent incident per the instructions in the support handbook; MarkLogic takes urgent incidents seriously, as every urgent issue results in a text message being sent to every support engineer, engineering management and the senior executive at MarkLogic.
- Enable Debug level logging so that any issues that arise can be more easily diagnosed. Debug level logging does not have any noticeable impact on system performance.