[Orca-users] Orca runs out of memory

[Michael Banker]

Yes, I think you're right.  It's running SuSE linux (SLES8, kernel 2.4.19 rev 340).  
We've been running it in daemon mode... it starts up using ~300gb, but keeps growing 
and I remember it approaching 2gb before it crashes.  


On Tue, 16 Dec 2003 rhfreeman at micron.com wrote:

> Do you know how much memory orca is using when it dies?
> 
> As it is running on x86 box, it is only 32-bit so can only handle 2GB of
> data memory by default.
> 
> You may wish to try another kernel. I've got Redhat 7.2 upto 2.7GB of
> data memory on a standard Redhat kernel.
> 
> I think there is a build option in the kernel to make this upto 3.5GB of
> memory, if you fancy building your own kernel.
> 
> Otherwise, split them out! :-) Or get a 64-bit machine to run it on!
> 
> Rich
> 
> > -----Original Message-----
> > From: orca-users-bounces+rhfreeman=micron.com at orcaware.com 
> > [mailto:orca-users-bounces+rhfreeman=micron.com at orcaware.com] 
> > On Behalf Of Michael Banker
> > Sent: Monday, December 15, 2003 9:27 PM
> > To: orca-users at orcaware.com
> > Subject: [Orca-users] Orca runs out of memory
> > 
> > 
> > Anyone else running Orca with large numbers of hosts?  We 
> > have thousands of servers running orcallator (solaris) and 
> > procollator (linux and aix) broken down into around 15 
> > groups.  One particular group has over 1,600 hosts.  At some 
> > point as the number of hosts reached this number, the orca 
> > server process started exiting without updating the 
> > index.html.  This is now consistently happening... it 
> > processes for 5-6 hours and just dies with an Out of Memory 
> > error (see stack trace below.)
> > 
> > This orca server runs on a Compaq DL580G2 with 16GB of ram, 
> > and 1GB swap.  When the process dies the server is not 
> > swapping, so that doesn't seem to be an issue.
> > 
> > The remaining groups are fine... the most hosts in any of the 
> > other groups is around 1,300, and there are two other groups 
> > with around 700 hosts.  All these groups are able to process 
> > the new data and update the index.html for that group.
> > 
> > Any ideas short of breaking down the larger group into smaller groups?
> > 
> > Thanks.
> > 
> > -- stack trace before proc exits --
> > mmap2(NULL, 2097152, PROT_NONE, 
> > MAP_PRIVATE|MAP_ANONYMOUS|MAP_NORESERVE, -1, 0) = -1 ENOMEM
> > (Cannot allocate memory)
> > mmap2(NULL, 1048576, PROT_NONE, 
> > MAP_PRIVATE|MAP_ANONYMOUS|MAP_NORESERVE, -1, 0) = -1 ENOMEM
> > (Cannot allocate memory)
> > mmap2(NULL, 4096, PROT_READ|PROT_WRITE, 
> > MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = -1 ENOMEM (Canno
> > t allocate memory)
> > brk(0x40000000)                         = 0x3ffff000
> > mmap2(NULL, 2097152, PROT_NONE, 
> > MAP_PRIVATE|MAP_ANONYMOUS|MAP_NORESERVE, -1, 0) = -1 ENOMEM
> > (Cannot allocate memory)
> > mmap2(NULL, 1048576, PROT_NONE, 
> > MAP_PRIVATE|MAP_ANONYMOUS|MAP_NORESERVE, -1, 0) = -1 ENOMEM
> > (Cannot allocate memory)
> > mmap2(NULL, 4096, PROT_READ|PROT_WRITE, 
> > MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = -1 ENOMEM (Canno
> > t allocate memory)
> > munmap(0xa9900000, 1048576)             = 0
> > munmap(0x40a17000, 524288)              = 0
> > munmap(0x583df000, 16384)               = 0
> > write(2, "Out of memory!\n", 15)        = 15
> > munmap(0xbff8f000, 4096)                = 0
> > munmap(0xbff8e000, 4096)                = 0
> > munmap(0xbff8c000, 4096)                = 0
> > munmap(0xbff8b000, 4096)                = 0
> > munmap(0xbff8a000, 4096)                = 0
> > munmap(0xbff89000, 4096)                = 0
> > munmap(0xbff88000, 4096)                = 0
> > 
> > 
> > -- 
> >   Michael Banker | Deutsche Bank | Parsippany, NJ | 973-606-3732
> > 
> > _______________________________________________
> > Orca-users mailing list
> > Orca-users at orcaware.com
> > http://www.orcaware.com/mailman/listinfo/orca-users
> > 
> 

-- 
  Michael Banker | Deutsche Bank | Parsippany, NJ | 973-606-3732