JIRA Data Center Troubleshooting

This page covers troubleshooting for a Data Center installation of JIRA. If you can't find an answer to your problem on this page, check out the JIRA Data Center FAQs.

Peer-to-peer connectivity

SymptomDisparity of data across nodes (e.g. email preferences not propagated)

Potential problemCaches not being synched across application nodes

Root cause Test Resolution
Misconfiguration: Multicast mis-configured Oracle multicast test, or omping Pass: All nodes can join the cluster
Fail: At least one node can't join the cluster
Reconfigure multicast
Misconfiguration: Firewall rules preventing propagation of multicast messages Check firewall logs Reconfigure firewall to allow multicast traffic
System design: Multicast traffic not on its own subnet (other network traffic burning up available bandwidth) Check configuration of all nodes Refactor network configuration to put multicast traffic on its own subnet

Symptom: Index not created on new node when it starts up

Potential problemNew node not getting a hot copy of index

Root cause Test Resolution
Database: Request for hot copy not received by other nodes due to faulty database connection on new node Verify that the request for the hot copy of the index was written to the database
i.e. Check clustermessage table on the db for "Index Backup" message and subsequent "Index backed up message"
 
Misconfiguration: One or more nodes not pointing to the correct location for the shared home directory Verify configuration of the shared home in the cluster.properties file  
Permissions: Existing nodes not able to write hot copy of index to shared home Verify that the existing nodes can write to the shared home:
i.e. Verify connectivity of shared home by creating a blank file on shared home from other node
Restore connectivity/efficacy of shared file system
Permissions: new node not able to read index from shared home. Verify that the new node can read from the shared home:
i.e. Check permissions
Change permissions to allow the application user to read files

Data inconsistency

SymptomUser not able to retrieve attachments associated with an issue

Potential problem: Attachments not being written to or read from the shared home for one or more nodes

Root cause Test Resolution
Permissions: One or more nodes not able to write to shared home directory Verify that the existing nodes can write to the shared home.
i.e. Verify connectivity of shared home by creating a blank file on shared home from other node
 
Misconfiguration: One or more nodes pointing to the wrong shared home directory Verify configuration of the shared home in the cluster.properties file Correct configuration
Check the configuration of the file share (possible misconfiguration) Correct configuration

Distribution of requests across nodes

SymptomRequests aren't being distributed across all nodes

Potential problemLoad balancer misconfiguration

Root cause Test Resolution
Misconfiguration: Load balancer not using the proper routing algorithm Check loadbalancer access log against application logs and verify loadbalancer configuration from administration interfac
i.e. Manually check with the browser each time and verify you are getting to other members of the pool by checking footer 
Correct configuration
Misconfiguration: IP addresses for one or more nodes not included in the range of IP addresses used by load balancer to route requests Enable access logging on Application server and verify traffic is reaching node
Pass: Traffic appears on the suspect node
Fail: No traffic appears on the suspect node
Correct routing


Potential problem
Connectivity issues with one or more nodes

Root cause Test Resolution
Node is unreachable from the loadbalancer Verify that DNS name for node is correct and verify that there is network connectivity
i.e. Telnet to the connector port of the node from the loadbalancer and check
Correct network

SymptomRequests aren't being distributed equally across all nodes

Potential problem: Load balancer misconfiguration

Root cause Test Resolution
Misconfiguration: Load balancer not using the proper routing algorithm Check loadbalancer access log against application logs and verify loadbalancer configuration from administration interfac
i.e. Manually check with the browser each time and verify you are getting to other members of the pool by checking footer 
Correct configuration


Potential problem
Users bypassing load balancer

Root cause Test Resolution
Configuration allows users to connect directly to each node Verify that an untrusted member cannot access the node Create filewall rules such that unauthorized users cannot connect to the node directly


Potential problem
Connectivity issues with one or more nodes

Root cause Test Resolution
     

SymptomRequests are not directed to the same node per client

Potential problemSessions aren't sticky

Root cause Test Resolution
Misconfiguration: Load balancer Troubleshooting load balancer stickiness Correct configuration

SymptomNodes are accessible and operational by the backdoor, but the base URL throws an error

Potential problemLoad balancer is down

Root cause Test Resolution
  Check for httpd processes running and check load balancer logs  

Potential problem: Load balancer misconfiguration

Root cause Test Resolution
The connection pool for the connector on Tomcat is lower than the number of requests kept open by the loadbalancer to that node Compare the connector configuration with the load balancer configuration, check for open connection numbers, and adjust the keepalive timeout.  
Incoming connection pool on the loadbalancer is exhausted grep for number of open connections and compare with the load balancer configuration  

Performance

SymptomHigh response time on one or more nodes

Potential problemNode is resource starved

Root cause Test Resolution
Process burning up resources on node (e.g. reporting, API calls) Check resourcing on individual nodes (db/cpu/iowait/swap) Allocate more resources to the individual node or configure loadbalancer to more fairly redirect traffic
Node server is underconfigured Evaluate server performance monitoring data to identify bottlenecks  

Potential problemNetwork link to node has bandwidth issues

Root cause Test Resolution
  Check network interfaces and connectivity between nodes Ensure that there are adequate interfaces for the traffic
Last modified on Sep 5, 2017

Was this helpful?

Yes
No
Provide feedback about this article

Not finding the help you need?

Ask the community

Powered by Confluence and Scroll Viewport.