Hello,
We use BI 4.2 SP3 WebI tool via launchpad. We’ve tremendous performance issue. I’m just a developer in the system but using this development environment sucks, really really sucks. It does not look like the administrator or systems architect who maintains the system really know what’s going on. I used to be BO administrator few years ago but never experienced this type of issue.
Here are the issues that I see in the system and I’d like to know what could cause these issues:
Log in to launchpad --normal
browsing to folder (any folder) --starts getting slow
Open report in View mode --slower (would give 5 out of 10)
Open report in Modify – really slow, takes over 6 minutes (would give 2 out of 10)
Report load time once opened --slowest, takes another 3-4 minutes (would give 0 out of 10 if possible)
The report contains about 5-10 tabs, clicking any tab takes over 3-5 minutes, just mind you that I’ve not refreshed this report yet – would give another 0 out of 10
clicking any object in the report or changing any object (like page number in footer), it takes over 3-5 minutes to get the report back to normal.
Once you log in to the launch pad, every clicks does something behind the scene and shows “Loading…” circle.
The “Refresh when open” check box is not checked in this or any other report.
By the time this report looks ok for you to work on then the 20 minutes time out get kicked off and you are out of the system
I’ve never worked as horrible as this before.
Nick,
Same report via Rich Client? Fine or the same?
They do not allow us to use any client tool. The only option we have is launchpad webi (no RC).
Marek,
Could be a problem with network connectivity too.
Agree, we really need smart network engineer to troubleshoot this issue. I can’t even issue basic commands like (ping/nslookup/netstat/telnet) from my machine, it’s a no no.
Also, is performance same or different when using the Java panel and the HTML panel?
Performance in both (Java/HTML) are the same.
What I know (as an extra information) is that this is out of the box installation, there is no additional configuration or setting change at all.
My goal is to list out all posibilities that’s bringing the environment to it’s knees and send an email to management to escalate it.
Here are my lists that could help to stabilize our current issue (as of today, would like to get your suggestion/additions):
BO:
Bind FRS/CMS/Processing to host/ip and specific port than leave it open to listen any host/port.
Add an additional processing servers
Run reposcan and fix findings or
3a. create a biar or promotion job and create a new CMS and release the biar/pm job to new CMS.
Restart SIA daily
App (Tomcat):
Bind Tomcat to a host/ip
Up the JVM from 2GB to 4GB (at least), these servers have 64GB of RAM.
Stop, clear cache, start Tomcat daily
Network:
Coordinate with network engineer for slowness, don’t have any good experience as most of them have no idea how to troubleshoot.
If it’s really a vanilla installation, meaning no further configuration and tuning has been performed, then it’s very likely that the default single APS is used for all servers. This is not a good idea and many of the performance issues you are experiencing could be caused by this. Have a look here where the APS split is described in more details:
(BOB member since 2002-08-15)
(BOB member since 2003-11-27)
to troubleshoot this issue. I can’t even issue basic commands like (ping/nslookup/netstat/telnet) from my machine, it’s a no no.
(BOB member since 2003-11-27)