On 4/9/2019 11:26 AM, mJg2XW wrote: > > Hi > Windows Server 2016
I had a client where that was happening and then it stopped. No good troubleshooting tools.
One thing we did was use jmap to dump the Java JVM and it was pretty small, only 150Megs, when the dhost.exe was at 12GB, so it was not 'IDM' at fault.
You can do this to test. find a JDK of the similar version to the one eDir/IDM is using. (cd to the jre/bin for eDir and try java -version then get a JDK that matches. Like 1.8 for 1.8. (Exact rev does not seem important).
Then get the PID (Process ID) from task manager for DHost.
Then run \path\to\jdk\jre\bin\jmap -dump:format=b,file=c:\temp\eDirDump.hpof 12345 (where 12345 is the PID number)
Then look at the size of the dump. It will at least rule out one possibility.
Then look in iMonitor at the Database cache you are set to for eDir on that box.
So if you have a 20GB dib and you are hard allocated (Not dynamic) to 15GB of cache, and you find eDir using all 15GB, that is not a bug, it is normal.
So add IDM's JVM contribution DIB cache and tell us how that compares to total memory usage.
I have one client who complains when DIB gets to 5GB but he has 512MB for JVM heap, and 4GB for DIB cache, so really .5GB is not a lot of memory leakage, as opposed to proper usage.
On 4/9/2019 12:17 PM, Lothar Haeger wrote: > Geoffrey Carman wrote: > >> I have one client who complains when DIB gets to 5GB but he has 512MB for JVM >> heap, and 4GB for DIB cache, so really .5GB is not a lot of memory leakage, >> as opposed to proper usage. > > ...and all the edir code and modules other than vrdim need to run in some > memory, too.
Which RPM version of the JDBC driver are you guys using, if any? I had a sad encounter with the one that came with 4.7 SP2, and then the 126.96.36.199 version (latest out there), both having pretty rotten bugs. The 188.8.131.52 (or whatever was just before 184.108.40.206) version seems to be working well enough, though.
The first version had constantly-growing state files; that was super-fun. The 220.127.116.11 version has a known issue where the driver constantly resync on the Publisher channel, causing the state files to constantly be rewritten.
I do not know that those would cause memory issues necessarily, but my customer was of course not on windows (SLES 12 SP4, eDir 9.1 SP3) so it's not exactly comparing apples to apples.
Restarting windows services once per day doesn't sound that unusual, does it? Sorry... I couldn't help it.
-- Good luck.
If you find this post helpful and are logged into the web interface, show your appreciation and click on the star below.
If you want to send me a private message, please let me know in the forum as I do not use the web interface often.
I have also this issue with one customer (IDM 4.6.2 on Windows 2012), we restart servers every day. At another customer the issue disappeared with eDir patch and IDM 4.6.2 and now still looks ok with IDM 4.7.2
mJg2XW <mJg2XW@no-mx.forums.microfocus.com> wrote: > Hi > > I'm only having SOAP and AD Driver running on the IDM 4.7.2/eDir 9.1.3 > > This memory bug where also in eDir 9.1.1 if I don't remember wrong. > > /Michael
We had memory leaks on windows 2916 with eDir 9.1.2. Since jumping to 9.1.3 it seems fine. Suspect it was a mama related leak.
Also saw memory issues on eDir 9.1.2 on RHEL 7.5. This was not resolved last I heard. Suspect it was audit/CEF related. However it became a lot less of an issue on a server without any auth activity. So could have also been NMAS.