LMS 2.0. At least once a week, there is a java process that pegs the processors at 100% and doesn't allow any other traffic to the server.
I have struck out as far as finding a solution. Anybody out there have a similar situation in the past?
Since we are going to collect flow information from the switch to catch all ICMP scanning activities,
please check if you can answer our question.
When a computer send out a ICMP request to another computer for only one time,
a flow entry would be created but the stat-bytes counter would be zero.
I know that this is normal since no subsequest packet use this flow entry again.
However, will this flow entry send out to the netflow collector ?
We found that only C6K sup2 can send out this information to the netflow collector.
Is it normal or a software bug ?
Switch> sh mls st en ip pro icmp
Destination IP Source IP Prot DstPrt SrcPrt Stat-Pkts Stat-Bytes
---------------- --------------- ----- ------ ------ ---------- ---------------
188.8.131.52 184.108.40.206 ICMP 0 0 0 0
220.127.116.11 18.104.22.168 ICMP 0 0 0 0
22.214.171.124 126.96.36.199 ICMP 0 0 0 0
188.8.131.52 184.108.40.206 ICMP 0 0 0 0
Hardware: C6K Sup1A
Result: Netflow collector cannot recevie ICMP type flow entries (with data witch zero count)
Hardware: C6K Sup2 (MSFC2)
Result: Netflow collector can receive ICMP type flow entries (with zero stat-byte count) successfully.
LMS 2.0 sometimes has problems with java when the server looses connetction to the LAN, perhaps this is the cause of your problem (if windows, look at the event console, if solaris you can see it in the syslog messages if LAN connection was down).
restart the management services:
(in a dos box): net stop crmdmgtd
wait 5 min
net start crmdmgtd
I have upgraded to LMS 2.2 and have all the latest patches and idu's installed but I still have the same problem. I have added memory to the server but has made no difference. Very annoying problem.
My processor is also regularily pinned at %100. java.exe is taking about %50 and cwjava.exe is taking the other %50.
I am running Windows 2000 Server w/ SP4 on a dual P3 600 MHz w/ 1GB of RAM. My LMS is 2.2 with all the latest updates and IDUs.
If you find a solution, please share.
which version of java do you use? After a reboot,without logging in into the application, does it have the same behavior (even after around 30 min?); how many devices are managed?
I have exactly the same kind of problems on LMS and ITEM plateforms upgraded with the latest version of CDone and IDUs. If you find a way to solve this issue, please share. Till know what we do is rebooting the servers to solve the problem.
Running into the same exact issue...with all of the latest patches as of 2 weeks ago. Did anyone ever come up with a fix for this javaw.exe 100% CPU issue?
I vaguely running into this problem after I put Update 1 on.
What version of java are you using, and do you have old versions that are running as well? If so, you will need to uninstall it.
We have the same problem. We have noticed the server will stay up for weeks with no problem if accessed by a client but the CPU becomes saturated in a few minutes with local access. Our workaroud is to access the information through a client not directly from the server.
Having had the same problem a while back, I eventually made a number of changes to resolve it. Unfornutely I cannot pin-point the problem solver as I made all the changes at once.
What I noticed is that when stopping the DISERVER and jrun proxy service the high CPU utilidies down.
Subsequently, I installed the new MDAC version 2.6. This made the selftest not fail on ODBC. I uninstalled Terminal Services on the server. Then used new databases from the orig folder. There might have been some other changes I've also done.
See if this helps.
Actually it turned out to be something totally out in left field. I only got the story second-hand, but...
It turns out there's some kind of bug where if you schedule a device inventory for more than 4 weeks out in the future, the scheduler in CW somehow rolls BACKWARD in time and gets all confused and the CPU just eats itself. By switching the device inventory to something more like a week or shorter, the problem mysteriously goes away.