[
https://issues.jboss.org/browse/TEIID-3596?page=com.atlassian.jira.plugin...
]
Steven Hawkins commented on TEIID-3596:
---------------------------------------
But isn't the fact that only next command can complete the log
for the previous one (when performed on the same client Statement) affecting the time
precision of the logs?
There are multiple notions of when a query ends. To be consistent we're using when
the request is closed on the server side. There is also when we are done producing
batches internally - however for large forward-only results that will always be correlated
to the results being read from the client as we do not indefinitely build up results on
the server side.
Or would they?
Any well designed client app should close resources as soon as possible. The timing also
shouldn't be used necessarily as an indication of server performance - that should be
measured from a client perspective. There is also TEIID-3591 to add a CPU utilization
measure to the command log as well.
Command logging: last command's log is not ended
------------------------------------------------
Key: TEIID-3596
URL:
https://issues.jboss.org/browse/TEIID-3596
Project: Teiid
Issue Type: Bug
Components: Query Engine
Affects Versions: 8.7.1.6_2
Reporter: Jan Stastny
Assignee: Steven Hawkins
There are incomplete entries when performing queries on some vdb and examining the
org.teiid.COMMAND_LOG entries in server.log.
From the logs it seems, that the closing entry for a user command is not logged until
either another command is performed or the server is shut down.
The following logs appears right after the command is performed:
{code:plain}
[org.teiid.COMMAND_LOG] (New I/O worker #64) START USER COMMAND:
startTime=2015-07-28 13:43:18.441 requestID=rfFDiTxhxLGL.0 txID=null
sessionID=rfFDiTxhxLGL applicationName=JDBC principal=user@teiid-security
vdbName=Portfolio vdbVersion=1 sql=select * from product
[org.teiid.COMMAND_LOG] (Worker1_QueryProcessorQueue1) START DATA SRC COMMAND:
startTime=2015-07-28 13:43:18.54 requestID=rfFDiTxhxLGL.0 sourceCommandID=0
executionID=0 txID=null modelName=Accounts translatorName=translator-h2
sessionID=rfFDiTxhxLGL principal=user@teiid-security sql=SELECT g_0.ID, g_0.SYMBOL,
g_0.COMPANY_NAME FROM Accounts.PRODUCT AS g_0
[org.teiid.COMMAND_LOG] (Worker0_QueryProcessorQueue2) END SRC COMMAND:
endTime=2015-07-28 13:43:18.548 requestID=rfFDiTxhxLGL.0 sourceCommandID=0
executionID=0 txID=null modelName=Accounts translatorName=translator-h2
sessionID=rfFDiTxhxLGL principal=user@teiid-security finalRowCount=25
{code}
while the following appers only when the server is shut down (notice the time difference
between this one and preceding ones):
{code:plain}
[org.teiid.COMMAND_LOG] (New I/O worker #64) CANCEL USER COMMAND:
endTime=2015-07-28 13:43:29.152 requestID=rfFDiTxhxLGL.0 txID=null
sessionID=rfFDiTxhxLGL principal=user@teiid-security vdbName=Portfolio
vdbVersion=1 finalRowCount=null
{code}
If there is another command following, the previous command log entry is completed (but
it can take even minutes):
{code:plain}
[org.teiid.COMMAND_LOG] (Worker0_QueryProcessorQueue3) END USER COMMAND:
endTime=2015-07-28 14:18:59.511 requestID=q97k+Z4gwV64.0 txID=null
sessionID=q97k+Z4gwV64 principal=user@teiid-security vdbName=Portfolio
vdbVersion=1 finalRowCount=25
{code}
Also the ending log entry of one command and opening entry of another appear in any
order.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)