RE: Problems in VM structure ?
I've adjusted MAXUSERS to 128 on my heavily loaded PIIs and the crashes have not re-occurred for 24 hours now. (Had to adjust NMBCLUSTERS up, though) The panics were happening every 5-8 hours like clockwork prior to this. I believe that these crashes are caused by heavy network traffic, not heavy load values, so a make world may not trigger this. Actually, I couldn't force it to happen when I hit the box hard during testing with web traffic, so it must be a combination thing. Another clue is the fact that I can't seem to get a Pentium (P5) to crash at all, ever, even when running exactly the same kernel config. The Pentium IIs fell over like crazy. -Troy Cobb Circle Net, Inc. http://www.circle.net To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: RE: Problems in VM structure ?
:I'm seeing different responses depending on hardware. : :On regular Pentium 166 machines, I almost NEVER get :a panic. On brand-new Pentium II 350s, I get a panic :every 6-9 hours. This happens when both kernels are :configured the same for maxusers. It happens when :both machines are under the same load level -- the :P5 stays rock solid, the P6 flakes out. : :What's the chance that our kernel adaptations for PIIs :is partly at fault? : :-Troy Cobb : Circle Net, Inc. : http://www.circle.net With what config? Have you tried reducing maxusers to 128? -Matt To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
RE: RE: Problems in VM structure ?
-Original Message- From: Matthew Dillon [mailto:dil...@apollo.backplane.com] :What's the chance that our kernel adaptations for PIIs :is partly at fault? : :-Troy Cobb : Circle Net, Inc. : http://www.circle.net With what config? Have you tried reducing maxusers to 128? -Matt I've had it at MAXUSERS=256 on both the P5 and the P6. The P5 stays stable, the P6 doesn't. If I reduce MAXUSERS to 128 then these heavily loaded boxen will fall over due to out of MBUFs errors, or so I believe. I'd love to find some real kernel-tuning documentation out there, one of my panics is a pipeinit: cannot allocate pipe -- out of kvm and I can't pull a crashdump due to a DSCHECK error because my SWAP is 2GB. -Troy Cobb Circle Net, Inc. http://www.circle.net To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: RE: RE: Problems in VM structure ?
Try reducing maxusers to 128. If you have mbuf problems, override NMBCLUSTERS ( making it 4096 or 8192 should be sufficient ). Sometimes network mbuf problems on heavily loaded machines are due to too-large default buffer sizes - if net.inet.tcp.sendspace or recvspace is greater then 16384, try reducing it to 16384. If your machine is not using that 2GB of swap, cap the swap at 1.9GB. -Matt Matthew Dillon dil...@backplane.com :I've had it at MAXUSERS=256 on both the P5 and the P6. The P5 stays :stable, the P6 doesn't. If I reduce MAXUSERS to 128 then these :heavily loaded boxen will fall over due to out of MBUFs errors, or :so I believe. : :I'd love to find some real kernel-tuning documentation out there, :one of my panics is a pipeinit: cannot allocate pipe -- out of kvm :and I can't pull a crashdump due to a DSCHECK error because my :SWAP is 2GB. : : :-Troy Cobb : Circle Net, Inc. : http://www.circle.net : : :To Unsubscribe: send mail to majord...@freebsd.org :with unsubscribe freebsd-current in the body of the message : To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: Problems in VM structure ?
[ CC trimmed ] tc...@staff.circle.net wrote in message ID a0cfa284c004d211b7ee0060082f32a412e...@freya.circle.net: I've had it at MAXUSERS=256 on both the P5 and the P6. The P5 stays stable, the P6 doesn't. If I reduce MAXUSERS to 128 then these heavily loaded boxen will fall over due to out of MBUFs errors, or so I believe. If you are running out of MBUF clusters, play with option NMBCLUSTERS=x directly rather than indirectly through maxusers... 4096 or 8192 is probably a good starting value. If you look in /sys/compile/YOUR_KERNEL at the param.c, then you can get a better idea of what maxusers tweaks, and what you need to tweak manually. I know its not a `howto' guide :) Sorry Gary -- Gary Palmer FreeBSD Core Team Member FreeBSD: Turning PC's into workstations. See http://www.FreeBSD.ORG/ for info To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: Problems in VM structure ?
On Tue, Feb 16, 1999 at 12:19:07AM -0800, Matthew Dillon wrote: :maxusers 256 Try reducing maxusers to 128. Another person reported similar behavior to me and after a bunch of work he tried going back to a basic distribution -- and everything started working again. To add some public informations. The host in case was using 512Meg RAM and I have tested it with 256Meg. I originaly installed a 3.0-CURRENT from mid December and updated to a recent version after getting panics. I used MAXUSERS of 512 and was able to trigger a panic during only a few minutes uptime after reducing it to 256 the host was more stable. Now it is running without any panics using MAXUSERS 128. It turned out that a maxusers value of 256 and 512 were causing his machine to go poof, but a maxusers value of 128 worked fine. I haven't tracked the problem down yet. Please try reducing your maxusers to 128 and email the results to current. -Matt To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message -- B.Walter To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: Problems in VM structure ?
On Wed, Feb 17, 1999 at 08:46:49AM -0500, tc...@staff.circle.net wrote: -Original Message- From: Matthew Dillon [mailto:dil...@apollo.backplane.com] :What's the chance that our kernel adaptations for PIIs :is partly at fault? : :-Troy Cobb : Circle Net, Inc. : http://www.circle.net With what config? Have you tried reducing maxusers to 128? -Matt I've had it at MAXUSERS=256 on both the P5 and the P6. The P5 stays stable, the P6 doesn't. If I reduce MAXUSERS to 128 then these heavily loaded boxen will fall over due to out of MBUFs errors, or so I believe. I my case it was a P-II 400 I'd love to find some real kernel-tuning documentation out there, one of my panics is a pipeinit: cannot allocate pipe -- out of kvm and I can't pull a crashdump due to a DSCHECK error because my SWAP is 2GB. The pipeinit is one of the panics I got very often - but they were not reduced to -Troy Cobb Circle Net, Inc. http://www.circle.net To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message -- B.Walter To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: Problems in VM structure ?
:maxusers 256 Try reducing maxusers to 128. Another person reported similar behavior to me and after a bunch of work he tried going back to a basic distribution -- and everything started working again. It turned out that a maxusers value of 256 and 512 were causing his machine to go poof, but a maxusers value of 128 worked fine. I haven't tracked the problem down yet. Please try reducing your maxusers to 128 and email the results to current. -Matt To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: Problems in VM structure ?
On Tue, 16 Feb 1999, Matthew Dillon wrote: :maxusers 256 Try reducing maxusers to 128. Another person reported similar behavior to me and after a bunch of work he tried going back to a basic distribution -- and everything started working again. It turned out that a maxusers value of 256 and 512 were causing his machine to go poof, but a maxusers value of 128 worked fine. I haven't tracked the problem down yet. Please try reducing your maxusers to 128 and email the results to current. For what it's worth, my maxusers is 250 and my system is quite stable, even during a make -j25 buildworld. -Matt To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message Brian Feldman_ __ ___ ___ ___ gr...@unixhelp.org _ __ ___ | _ ) __| \ http://www.freebsd.org/ _ __ ___ | _ \__ \ |) | FreeBSD: The Power to Serve! _ __ ___ _ |___/___/___/ To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: Problems in VM structure ?
On Tue, 16 Feb 1999, Matthew Dillon wrote: MD Try reducing maxusers to 128. Another person reported similar behavior MD to me and after a bunch of work he tried going back to a basic MD distribution -- and everything started working again. Hmmm, ok. MD It turned out that a maxusers value of 256 and 512 were causing his machine MD to go poof, but a maxusers value of 128 worked fine. Ok. I'm glad, in a way, that I'm not the only one seeing this. The really weird thing though is that since reporting the problem, it hasn't re-occured. If it occurs again, I'll mail the results of the gdb -core /var/crash/blah, a trace and then try reducing the number of maxusers. This is the longest uptime I've had in almost two weeks - 14 hours. Here's hoping :) MD I haven't tracked the problem down yet. Please try reducing your maxusers MD to 128 and email the results to current. If the problem re-occurs, I'll do so :) --- Khetan Gajjar (!kg1779) * khe...@iafrica.com ; khe...@os.org.za http://www.os.org.za/~khetan * Talk/Finger khe...@chain.freebsd.os.org.za FreeBSD enthusiast* http://www2.za.freebsd.org/ Security-wise, NT is a OS with a kick me sign taped to it To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: Problems in VM structure ?
On Tue, 16 Feb 1999, Matthew Dillon wrote: Try reducing maxusers to 128. Another person reported similar behavior to me and after a bunch of work he tried going back to a basic distribution -- and everything started working again. It turned out that a maxusers value of 256 and 512 were causing his machine to go poof, but a maxusers value of 128 worked fine. Another datapoint, Sybase goes poof with maxusers set to 64 or higher. This has been the case since before 3.0 was released. -john To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: Problems in VM structure ?
Matthew Dillon said: :maxusers 256 Try reducing maxusers to 128. Another person reported similar behavior to me and after a bunch of work he tried going back to a basic distribution -- and everything started working again. It turned out that a maxusers value of 256 and 512 were causing his machine to go poof, but a maxusers value of 128 worked fine. I haven't tracked the problem down yet. Please try reducing your maxusers to 128 and email the results to current. Likely because data structures are getting too big. The kernel is limited to (I forget) how big in VA space. -- John | Never try to teach a pig to sing, dy...@iquest.net | it makes one look stupid jdy...@nc.com | and it irritates the pig. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
RE: Problems in VM structure ?
I'm seeing different responses depending on hardware. On regular Pentium 166 machines, I almost NEVER get a panic. On brand-new Pentium II 350s, I get a panic every 6-9 hours. This happens when both kernels are configured the same for maxusers. It happens when both machines are under the same load level -- the P5 stays rock solid, the P6 flakes out. What's the chance that our kernel adaptations for PIIs is partly at fault? -Troy Cobb Circle Net, Inc. http://www.circle.net -Original Message- From: Brian Feldman [mailto:gr...@unixhelp.org] Sent: Tuesday, February 16, 1999 7:48 AM To: Matthew Dillon Cc: Khetan Gajjar; curr...@freebsd.org Subject: Re: Problems in VM structure ? On Tue, 16 Feb 1999, Matthew Dillon wrote: :maxusers 256 Try reducing maxusers to 128. Another person reported similar behavior to me and after a bunch of work he tried going back to a basic distribution -- and everything started working again. It turned out that a maxusers value of 256 and 512 were causing his machine to go poof, but a maxusers value of 128 worked fine. I haven't tracked the problem down yet. Please try reducing your maxusers to 128 and email the results to current. For what it's worth, my maxusers is 250 and my system is quite stable, even during a make -j25 buildworld. -Matt To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message Brian Feldman_ __ ___ ___ ___ gr...@unixhelp.org _ __ ___ | _ ) __| \ http://www.freebsd.org/ _ __ ___ | _ \__ \ |) | FreeBSD: The Power to Serve! _ __ ___ _ |___/___/___/ To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: Problems in VM structure ?
Hi. I saw that my 4-CURRENT box from 8 February dropped to ddb after my last make world. I rebuilt world today, and the same problem is occuring. These problems started occuring after Matt Dillon's changes to the VM system. What is worrying/troubling is that in single user mode, the machine is stable, and manages to build world without a problem. When booted into multi-user mode, it's stable and usable for anywhere from 1 to 3 hours, and then panics. There are no active users on at the time, and the machine is not heavily loaded (0.0-0.2) I suspected a hardware error, so swopped all the RAM from a production machine, and it still produces the same fault. The error is panic: vm_fault: fault on nofault entry, addr : f2572000 This indicates an unmapped struct buf, should be a software bug. Debugger (panic) Stopped at Debuger+0x37: movl $0,in_Debugger When I hit c, I get this : Could you type in bt next time this happens, and post the result? -lq To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: Problems in VM structure ?
On Monday, 15 February 1999 at 18:00:16 -0500, Luoqi Chen wrote: Hi. I saw that my 4-CURRENT box from 8 February dropped to ddb after my last make world. I rebuilt world today, and the same problem is occuring. These problems started occuring after Matt Dillon's changes to the VM system. What is worrying/troubling is that in single user mode, the machine is stable, and manages to build world without a problem. When booted into multi-user mode, it's stable and usable for anywhere from 1 to 3 hours, and then panics. There are no active users on at the time, and the machine is not heavily loaded (0.0-0.2) I suspected a hardware error, so swopped all the RAM from a production machine, and it still produces the same fault. The error is panic: vm_fault: fault on nofault entry, addr : f2572000 This indicates an unmapped struct buf, should be a software bug. Debugger (panic) Stopped at Debuger+0x37: movl $0,in_Debugger When I hit c, I get this : Could you type in bt next time this happens, and post the result? It's t in ddb, not bt. Isn't consistency wonderful? Khetan, you should also take a dump. The backtrace is a good start, but it probably won't be enough to solve the problem. Greg -- See complete headers for address, home page and phone numbers finger g...@lemis.com for PGP public key To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message