Hello! Unfortunately, 2.4 -aa VM is not good too. Personally I never had problems with 2.4 VM, but we run Oracle for not more than 30 users, but I see that there are VM problems with 2.4.18 kernels (kswap eats too much cpu), this is why now I test rmap VM. But I know guys which wanted to run Oracle for about 700 users on Linux. They tried RH AS 2.1, RH 7.x, SLES7 and finally decided that Linux can't handle more then about 400 Oracle users (well, w2k limit is about 350 ;-) ). So, now they run Tru64 :-( But I know guy which run Oracle for about 700 users on Unixware, so this is not x86 hardware problem :-) As you can see here http://lists.suse.com/archive/suse-oracle/2002-Jul/0266.html http://lists.suse.com/archive/suse-oracle/2002-Jul/0267.html there are people which run Oracle for 500 users on Linux, but only using 2.2 kernel. If you want Linux, you can try 2.2 kernel :-) You have only 4Gb RAM, 2.2 can use up to 4Gb of RAM, so I don't see any problems. Or, alternatively, you can go to Unixware (OpenUnix) - www.sco.com
----- Original Message ----- From: "Andy Beal" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> Sent: Thursday, December 12, 2002 5:12 PM Subject: Re: slow performance on RH advanced Srv 2.1 > Hello, > > I'm also running Advanced Server 2.1, except this system is running > Oracle 9i. > I've talked to Redhat several times and been extremely disappointed > with their > answers to this problem. It seems to have gotten better with patches, > but it's still > a big issue for me. They say that engineering it working on it. One > solution they > suggested (which is not possible due to oracle memory requirements in > my situation), > was to remove some memory, I have 4GB Ram. It's nice to hear you've > got the same > issue with 2Gb. > > I seem to have the best luck with the kernel I'm running now: > 2.4.9-e.10 Enterprise > > The SMP kernel seems to misbehave more than the Enterprise kernel. > > I am still hanging on looking for the patch that fixes it, however each > day > I get closer and closer to ditching their kernel enhancements and > building my > own kernel. > > I call them every two weeks and report the problem again, hoping they > will notice that > it is a problem. If anyone else is experiencing this issue, perhaps we > can get > them to move faster. > > Andy > > >>> [EMAIL PROTECTED] Saturday, December 07, 2002 5:18:50 AM >>> > Hye, > > we are using RH AdvSrv 2.1 as a HA cluster for filesharing (samba, > nfs) > on Dell 1650 (2x1133MHz, 2 GB RAM for each server). We share a 400GB > diskunit as filespace. > > We detected a high cpu-usage in addition with low performance on the > clients while accessing the fileshares. > > We think kswap is the reason, because this process consums most > cputime > and sometimes it uses around 90% cpu-usage. > > Is there someone who has the same effect? Are there any known patches > for this problem? > > I read a lot about kswapd/kupdated and problems in kernels between > 2.4.3 > and 2.4.15. The kernelmailinglist tells me that this was fixed in > 2.4.16 > standard kernel. > The problem is that redhat uses a highly patched kernel and i'm not > sure > if i can uses a standardkernel distribution and where to get the > fixes. > > Thanks for all ideas/infos to this issue. > > regards > > Mesiol > > > > > -- > redhat-list mailing list > unsubscribe mailto:[EMAIL PROTECTED]?subject=unsubscribe > https://listman.redhat.com/mailman/listinfo/redhat-list > > > > -- > redhat-list mailing list > unsubscribe mailto:[EMAIL PROTECTED]?subject=unsubscribe > https://listman.redhat.com/mailman/listinfo/redhat-list > -- redhat-list mailing list unsubscribe mailto:[EMAIL PROTECTED]?subject=unsubscribe https://listman.redhat.com/mailman/listinfo/redhat-list
