This post is more than 5 years old
7 Posts
0
155626
6.5 Upgrade, now poor performance
Hi,
Let me preface by saying I also am opening a support case. However looking for suggestions on this one. Prior to vFoglight 6.5 upgrade pages were loading in about 5 seconds. Now logging in can take upwards of 180 seconds and various pages timeout regularly. We have the following setup:
Using VM with 4 vCPU (2.6 GHz AMD six cores underlying) 16 GB RAM (no swapping, ballooning, or such), using MSSQL 2005 backend. Current tuning params in foglight.config are:
foglight.vm.option0="-Xmx12g";
foglight.vm.option1="-Xms12g";
foglight.vm.option2="-XX:+DisableExplicitGC";
foglight.vm.option3="-Dfoglight.persistence.cache.duration=30000000";
We are monitoring 5 vCenters, about 2000 VMs. This is not a simple environment, however according to the sizing guidance our VM should be adequate. Any suggestions on where to look to track down the problem or additional tuning thoughts?
Thanks in advance.
Let me preface by saying I also am opening a support case. However looking for suggestions on this one. Prior to vFoglight 6.5 upgrade pages were loading in about 5 seconds. Now logging in can take upwards of 180 seconds and various pages timeout regularly. We have the following setup:
Using VM with 4 vCPU (2.6 GHz AMD six cores underlying) 16 GB RAM (no swapping, ballooning, or such), using MSSQL 2005 backend. Current tuning params in foglight.config are:
foglight.vm.option0="-Xmx12g";
foglight.vm.option1="-Xms12g";
foglight.vm.option2="-XX:+DisableExplicitGC";
foglight.vm.option3="-Dfoglight.persistence.cache.duration=30000000";
We are monitoring 5 vCenters, about 2000 VMs. This is not a simple environment, however according to the sizing guidance our VM should be adequate. Any suggestions on where to look to track down the problem or additional tuning thoughts?
Thanks in advance.
mcondy
18 Posts
0
December 9th, 2010 11:00
I don't know if you have an open support case for this but the way you discuss I suspect that you may. We have spotted a couple of issues that affect larger implementations and should be able to help. please email me the case number(s) (mike.condy@quest.com so that I can have the R&D team direct the support engineer accordingly.
Regards
-Mike C
lmurphy1
57 Posts
0
December 6th, 2010 15:00
Your heap size is too high. 12GB is a lot to assign to the heap and is probably spending a lot of time in garbage collection.
I would not go higher than 6GB on the heap size unless instructed by support.
You also can take out option 2 and 3 as these are set already by default and are just redundant.
What pages are timing out? Support will need to know what specific dashboards are timing out.
Does it take 180 seconds for all users to login? Or is it just certain users? If certain users, what is there homepage set as?
-Larry
thejimmy1
7 Posts
0
December 6th, 2010 16:00
As far as pages timing out, it really can be any of them. On the VMware enviornment dashboard for example the past 4 hours view will load, change it to 72 hours and it times out. Every user is using the "perspective" new welcome page as far as I'm aware. 3 users including the super user foglight we've tried and all either timeout or get in eventually. We have not been able to find a user with a changed homepage to test with. Since our foglight user is also suffering, we can't get in to change someone's homepage either.
We are going to try removing all the config options and see if it helps. It's also curious that they don't use large pages on 64 bit since the new JVM supports it.
Thanks.
thejimmy1
7 Posts
0
December 6th, 2010 21:00
thejimmy1
7 Posts
0
December 9th, 2010 04:00
Anyone have similar issues? At this point our FMS has become the bain of my existence since I've spent 2.5 days troubleshooting.
mrvirtual
3 Posts
0
December 21st, 2010 01:00
thejimmy1
7 Posts
0
January 24th, 2011 21:00
Here are some sample errors that seem to appear when hitting certain pages:
2010-12-05 12:56:07.529 WARN [Data-3-thread-109] com.quest.nitro.service.rule.rule.Rule - Failed to evaluate expression qObjs = #capacityAvailable from $scope.hostLogicalDisk#;
availbility = qObjs?.values(qObjs?.topologyObjects[0])?.get(0)?.getValue().getAvg();
return availbility; in rule: VMW Virtual Machine Logical Drive Utilization. Reason: com.quest.nitro.service.sl.interfaces.scripting.ScriptAbortException: script1000147: java.lang.IndexOutOfBoundsException: Index: 0, Size: 0
Here are the options we added to the config:
foglight.vm.option0="-Xmx6g";
foglight.vm.option1="-Xms6g";
foglight.vm.option2 = -"XX:PermSize=256m";
foglight.vm.option3 = "-XX:MaxPermSize=256m";
foglight.vm.option4 = "-XX:NewRatio=5";
foglight.vm.option5 = "-Dsun.rmi.dgc.client.gcInterval=3600000";
foglight.vm.option6 = "-Dsun.rmi.dgc.server.gcInterval=3600000";
foglight.vm.option7 = "-XX:LargePageSizeInBytes=256m";
foglight.vm.option8 = "-XX:+DisableExplicitGC";