Slow Performance of Mobility during peak hours

Document ID:7009812
Creation Date:28-Nov-2011
Modified Date:21-Apr-2015
- Micro Focus Products:
  GroupWise
  GroupWise Mobility Service

Environment

SUSE Linux Enterprise Server 11

Novell Data Synchronizer Mobility Pack

Situation

Slow performance of Mobility Server during peak hours

Mails reach on the Device after 45 minutes

Mails does not reach the Device immediately

Mobility works fine during non-peak hours

Mails and Appointments reach the device after a long time during working hours

Conversion queue in /var/log/datasync/connectors/default.pipeline.mobility-Appinterface.log keeps on increasing and takes a long time to convert
WebAdmin shows The disk is experiencing a high work load. Disk busy.

Resolution

This can happen if there is a bottle neck at the Input/Output layer. Please run tsatest to make sure that the Effective MB/Min is atleast around 800 MB/min. Please follow the steps listed below:

Copy the novell-sms rpm from GroupWise SoftWare Distribution Directory\agents\linux to the mobility server.
Type "rpm -ivh novell-sms-1.1.4-16.i586.rpm " (Replace novell-sms-1.1.4-16.i586.rpm with the exact name of the rpm)
Once it is installed, please type "rcnovell-smdrd start " and press Enter.
Shut down mobility by typing "rcdatasync stop " and press Enter.
Type "/opt/novell/sms/bin/tsatest --path /var/ "
Enter the Username as root and the root's password.
Watch the Effective MB/min and it is recommended that it should be atleast around 800 MB/min.

Note: If it is lower than that, then please fine make sure that /var is moved to a disk with faster Input/Output or fine tune the Disk if possible.

Also, please follow the steps listed below to disable the Access Time attribute for files as it is not needed for Mobility. Please be careful while following the steps listed below as incorrect modifying the file may cause the server not to boot up:

Edit /etc/fstab .
Add the noatime attribute for /var partition as follows:
/dev/sda5 /var ext3 acl,user_xattr,noatime 1 2
Save the file and restart the Server.

Also, if this the Linux Server is on VMware:

Please make sure VMware Tools are installed and running. Please type /etc/init.d/vmware-tools status to check if VMware Tools are running. If VMware Tools are not running, please follow the VMware documentation to install VMware Tools.
Consider configuring the I/O Scheduler to use NOOP (noop): See Linux 2.6 kernel-based virtual machines experience slow disk I/O performance (2011861).

Additional Information

Conversion queue in /var/log/datasync/connectors/default.pipeline.mobility-Appinterface.log can be checked by grepping for ": processing "

mymobile:/var/log/datasync/connectors # tail -f default.pipeline1.mobility-AppInterface.log | grep ": processing"

2011-11-28 15:08:12.119 DEBUG_VERBOSE [Thread-4] [syncEngineEventMonitor:49] [userID:] [eventID:] [objectID:] [Monitor] SyncEngineEvents: processing 40 of 120
2011-11-28 15:41:39.292 DEBUG_VERBOSE [Thread-4] [syncEngineEventMonitor:49] [userID:] [eventID:] [objectID:] [Monitor] SyncEngineEvents: processing 40 of 150