GDC: "Connection Refused", "Failed to retrieve saved searches", "Failed to load existing record sets" and other error messages
- Product: Voyager
- Product Version: Voyager 9.0
- Relevant for Installation Type: select the relevant: Dedicated-Direct, Direct, Local, Total Care
Symptoms
Attempting to use GDC results in any of the following error messages although GDC and network configurations appear to be correct:
- When attempting to log into GDC:
- "Connection refused"
- "Connection failed"
- While attempting to run jobs or access data in GDC after logging in:
- "Failed to retrieve saved searches"
- "Failed to load existing record sets"
- "Failed to load rule sets"
- "Failed to load rule set groups"
- "Failed to submit data change job. Error code = 1"
- Additional error messages during GDC use are possible.
- Restarting jobd and gdcsvr appears to temporarily resolve the issue, however the issue recurrs within a few days.
- jobd and gdcsvr appear to be running on the server but also appear to be unresponsive.
- No report of the issue appears obvious in any system or application log.
Cause
- These errors can occur if GDC needs to be restarted on the server.
- These errors can also occur if the memory allotment for the jobd and gdcsvr processes is insufficient.
Resolution
IMPORTANT: Before going through the "Add Memory" procedure described below, kill any GDC processes that are running and then restart GDC as per the instructions found in the following article, and then retest GDC to see if the issue is resolved:
How to find out whether Voyager Global Data Change (GDC) is running
Otherwise:
Contact Voyager Support (login required) to request assistance in making the following change; OR use the "Add Memory" procedure found below to increase the Java heap size for gdcsvr and jobd.
Procedure to Add Memory for gdcsvr and jobd
The following procedure assumes a basic understanding of the linux/unix command line. Ex Libris recommends always backing up configuration files before making changes. Please contact Voyager support for assistance if you are uncomfortable with the following procedure or are unable to access your server:
- Log into the server as the voyager user and for the affected instance, open /m1/voyager/xxxdb/ini/voyager.env.
- At the bottom of the file locate the following two lines:
GDCSVR_JAVA_OPTS="-Xms16m -Xmx32m -XX:+UseConcMarkSweepGC -Dfile.encoding=UTF-8"
JOBD_JAVA_OPTS="-Xms16m -Xmx32m -XX:+UseConcMarkSweepGC -Dfile.encoding=UTF-8"
- Increase the Java heap size for each process by increasing the values for -Xms and -Xmx to 64m OR double the previous value, whichever is greater.
- If the value for -Xmx is greater than -Xms, make the adjustment based on the higher valued and increase -Xms to match -Xmx:
GDCSVR_JAVA_OPTS="-Xms64m -Xmx64m -XX:+UseConcMarkSweepGC -Dfile.encoding=UTF-8"
JOBD_JAVA_OPTS="-Xms64m -Xmx64m -XX:+UseConcMarkSweepGC -Dfile.encoding=UTF-8"
- Run: ps -efw | egrep "jobd|gdc"
to check for running gdcs and jobd processes for the appropriate instance. Make note of the PID of any running processes. ( Note: -w option prevents ps -ef from truncating output to the width of the terminal window, which may prevent grep from correctly spotting the server processes.) - Run "kill <PID>" for each runng gdc or jobd process if necessary.
- Restart jobd and gdcsvr by running "/m1/voyager/xxxdb/sbin/Pjobd" and "/m1/voyager/xxxdb/sbin/Pgdcsvr"
- Retest GDC.
- Article last edited: 08-Oct-2020