CPS 12 sqlservr.exe is sending 100 MB/Sec to the Java.exe of the app server? That's about 800 mbits/sec in a 1 gbits switch. what's going on?
anyone know whats' "useoneconnection=0" means in EMR.ini file?
Don't know anything about that but I was told that in CPS 12/EMR 9.8 they had moved a lot of functionality from the client to the JBoss server and that it was going to need much more memory and processor than before.
First it was killing the network traffic (800 mbits/sec). Now CPU is at 100%. Memory not an issue.
DavidShower said:
Don't know anything about that but I was told that in CPS 12/EMR 9.8 they had moved a lot of functionality from the client to the JBoss server and that it was going to need much more memory and processor than before.
This is our understanding as well. While it can only improve over time, I suppose, we now have the worst of both worlds with insanely high client AND server system requirements.
OP: I'm curious about your site size. We're in the 350-400 concurrent user range.
We are a much small site about 140 concurrent users. I restarted jboss and it's back to normal but from time to time, the CPU goes up 100% and stay there until jboss get restarted again.
lic08 said:
We are a much small site about 140 concurrent users. I restarted jboss and it's back to normal but from time to time, the CPU goes up 100% and stay there until jboss get restarted again.
I restart jboss nightly.
The java.exe is taking up 100% of the cpu and for the past few days, it occurred only at night. After i restart jboss it calms down and runs normally throughout the business hours. Does the app server runs anything at night that would cause this? and if it is safe to restart the jboss nightly? Thanks.
I recommend calling GE. They have "Infinite Wisdom (tm)" regarding the Centricity operating environment.
If you have a "small" environment, I have a teeny environment. I do a cold backup of the db every night starting at 12:20 AM. Before that time, I stop the DTS, jboss, and all other dependent processes that I've learned are not so robust when the db goes dark. After restarting the db, the other services are restarted.
These stops/starts are done via timed jobs, and on different servers.
It's messy, but my job is to have services up and running, without fail, when the office opens.
I'm working with two GE analylists on this issue already. When GE fails, CHUG is a very good source.
lic08 said:
The java.exe is taking up 100% of the cpu and for the past few days, it occurred only at night. After i restart jboss it calms down and runs normally throughout the business hours. Does the app server runs anything at night that would cause this? and if it is safe to restart the jboss nightly? Thanks.
If you're an ePrescribing user, you can check the eSM console settings to see if the 100% spike coincides with the nightly formulary job.
Well i can't tell. The formulary process takes place any time after 7 pm but the spikes occurred at 9 pm.
Network traffic is back to 800 mb/secs and this time restarting the jboss does not help. In fact, restarting all the servers didn't work either. I'm trying to reinstall jboss and i can't even get past the serversetup->utilities db login. I get" error when creating folder on remote server. error details: logon failure. the target account name is incorrect." This has to be the worst cps upgrade.
Problem solved. After talking to 4-5 GE supports, finally one told me that it is because of CQR and Qvera. It's an known issue. After stopping CQR and QVera, everything drop back to normal. No need to restart jboss. As for CQR and Qvera, it's back to engineering.
Thanks so much for posting this and letting us know. Are you just bouncing (restarting) CQR and Qvera to clear the traffic storm?