Re: Cluster cdbd service process

Thats another limit i think, has nothing to to with the cdbd service
process... 

this was a limit of the drdb driver, because you can't compile the
kernel with highmem support and use the drdb driver...


but i posted yesterday patches for cdbd to use failsafe under RH 7.2,
SuSE 7.2 or other glibc 2.2.4 based distro. there was a bug in the first
version of the patches leaving a lot of sockets in state close_wait
around. 


oliver


On Thu, 2002-04-25 at 17:59, Austin Gonyou wrote:
> So are you saying for say something like Oracle failover, or LVM
> failover, I'd be limited to 896MB?
> 
> On Wed, 2002-04-24 at 00:05, Oliver Jehle wrote:
> > When you sum all failsafe processes using the cdbd (database handler),
> > then you will see that the amount of cdbd processes is related to. start
> > a manangement client, the amount will increase... 
> > 
> > every process using the cluster database uses rpc and then cdbd handles
> > the connnection in a thread. in linux, threads are implemented as
> > processes . so you will have a lot of cdbd processes..
> > 
> > 
> > 
> > On Wed, 2002-04-17 at 14:34, Christoph Biardzki wrote:
> > > 
> > > Hi,
> > > 
> > > some time ago I noticed that the number of cdbd processes in my cluster
> > > is quite high (around 30) - the cluster works fine: but - is this normal?
> > > 
> > > Cheers,
> > > 
> > > Christoph
> > > 
> > > 
> > > bash-2.05# pstree
> > > init-+-atd
> > >      |-bdflush
> > >      |-cdbd---cdbd---32*[cdbd] <== ************
> > >      |-cmond-+-cad---cad---9*[cad]
> > >      |       |-crsd---crsd---crsd
> > >      |       |-ha_cmsd---ha_cmsd---ha_cmsd
> > >      |       |-ha_fsd
> > >      |       |-ha_gcd
> > >      |       |-ha_ifd
> > >      |       `-ha_srmd---ha_srmd---2*[ha_srmd]
> > >      |-cron
> > > 
> > > 
> > > _______________________________________________
> > > LinuxFailSafe mailing list
> > > [email protected]
> > > http://lists.community.tummy.com/mailman/listinfo/linuxfailsafe
> > 
> > 
> > _______________________________________________
> > LinuxFailSafe mailing list
> > [email protected]
> > http://lists.community.tummy.com/mailman/listinfo/linuxfailsafe
> -- 
> Austin Gonyou
> Systems Architect, CCNA
> Coremetrics, Inc.
> Phone: 512-698-7250
> email: [email protected]
> 
> "It is the part of a good shepherd to shear his flock, not to skin it."
> Latin Proverb