nxshell - Database failure on eventlog query

[zev]

Check your server log on debug level 4. There should be something like: "LOG QUERY: %s"

[zshnet]

I looked at it once but I can't seem to remember where the log file should be located on Linux. Any suggestions as to where I could look?

Thanks,
Zach

[zev]

Check "LogFile" parameter in server config(NETXMSHOME/etc/netxmsd.conf or /etc/netxmsd.conf). If it is not defined try to check something like /var/log/netxmsd.

[zev]

Optimal log level for debugging problems is 6.

[zshnet]

Good to know. We're currently running really hard on Debug levels, I think. The only relevant line I could find in any of the five (netxmsd.0/1/2/3/4) logfiles was this:
[20-Jun-2016 12:31:54.204] [DEBUG] [CLSN-4] User [email protected] authenticated (language=en clientInfo="nxjclient/2.0.3 (Linux 3.19.0-58-generic; libnxcl 2.0.3)")

Our debug level is currently at 6. We have a million of these messages:
[20-Jun-2016 12:31:54.245] [DEBUG] [CLSN-4] Sending message CMD_OBJECT

which is making it a bit hard to read through the logs. Searching for "database" "nxc" or similar terms got me nothing. Any guesses? Should I try to dig into the postgresql logs?

Thanks,
Zach

[zshnet]

That was it! I looked in the postgresql log:

2016-06-20 12:37:31 EDT ERROR:  column "source" does not exist at character 159
2016-06-20 12:37:31 EDT STATEMENT:  SELECT event_id,event_timestamp,event_source,dci_id,event_code,event_severity,event_message,user_tag,root_event_id FROM event_log WHERE event_id<=934558 AND (Source BETWEEN 0 AND 8153) LIMIT 1000

I guess the column isn't named Source? I'll have to look in the postgresql database to see what's there, unless you have a better idea of the column name.

Thanks,
Zach

[zev]

The name is "event_source". UI is hiding column real names. You should use names that are defined in database.

[zshnet]

One last question, I hope! It appears to not be filtering anything. Running the following code, I get five events, 10, 11, 12, 13, 109. The filter should have spat out 27628, right? If I pull the columnfilter from the logfilter, I get that the filter is still in place, but I don't seem to get any filtering. Any guesses?

eventLog = session.openServerLog("EventLog")

columnFilter = ColumnFilter(ColumnFilter.EQUALS,27628)
columnFilter.setNumericValue(27628)
logFilter = LogFilter()
logFilter.setColumnFilter("event_id",columnFilter)
#print(columnFilter.getNumericValue())
#print(columnFilter.getOperation())
fill = logFilter.getColumnFilter("event_id")
print(fill.getNumericValue())
eventLog.query(logFilter)

table = eventLog.retrieveData(0,5)
for j in range(table.getRowCount()):
    #for i in range(table.getColumnCount()):
    print(table.getColumnDisplayName(0))
    print(table.getCellValue(j,0))

output is:
27628 [output of fill.getNumericValue()]
event_id
10
event_id
11
event_id
12
event_id
13
event_id
109

[zshnet]

For some quick information, I chose this event because it was not filtering on event_severity, so I gave ID a try, and it seems to similarly not work. I have pasted the event below, so that you can see it exists.

27628 19.05.2016 11:56:32 Karmaloop-R1 0 SYS_NODE_CRITICAL Critical Node status changed to CRITICAL 0

[zev]

I found problem. It's bad class declaration. You want function "public ColumnFilter(int type, long value)" to be used, but both values are counted like long and class is created like: "public ColumnFilter(long rangeFrom, long rangeTo)". So your values were counted like a range. Type should be change to Enum type. I have created bug on it: https://dev.raden.solutions/issues/1245.
Temporary solution:
import java.lang.Integer
eventLog = session.openServerLog("EventLog")

columnFilter = ColumnFilter(java.lang.Integer(ColumnFilter.EQUALS),363)
logFilter = LogFilter()
logFilter.setColumnFilter("event_id",columnFilter)
...

[zshnet]

Hey, thanks! That works perfectly. Thanks again for all your work + help!

Have a good one,
Zach