Re: Kernel options help
On Thu, Apr 14, 2011 at 10:00 AM, Bruce Dubbs bruce.du...@gmail.com wrote: There is definitely something wrong. On a production LFS system running in a virtual envronment, I get: real 0m18.514s user 0m8.984s sys 0m2.697s I'm still having the same problems after recompiling the kernel a few times with different options. Ken, I don't know how to check against the 32 v 64 bit. uname reports: Linux hojo-lfs-6.8 2.6.38.2-LFS6.8 #1 SMP Mon Apr 18 20:35:16 MDT 2011 x86_64 x86_64 x86_64 GNU/Linux I have picked out a shell script that runs very slowly from the configure script from openssh. On a stable system the script runs as fast as expected: Script: #LS nuisances. for as_var in \ LANG LANGUAGE LC_ADDRESS LC_ALL LC_COLLATE LC_CTYPE LC_IDENTIFICATION \ LC_MEASUREMENT LC_MESSAGES LC_MONETARY LC_NAME LC_NUMERIC LC_PAPER \ LC_TELEPHONE LC_TIME do if (set +x; test -z `(eval $as_var=C; export $as_var) 21`); then eval $as_var=C; export $as_var else ($as_unset $as_var) /dev/null 21 $as_unset $as_var fi done The time on the script my box: real0m5.780s user0m0.030s sys 0m5.710s I have to run this script in a loop roughly 300 times on the stable box to approximate the LFS run. If I run this in a loop on the LFS box top shows the following while bash only consumes around 7 to 9% of the cpu and a scant of memory. Cpu(s): 0.0%us, 25.7%sy, 0.2%ni, 74.0%id, 0.0%wa, 0.0%hi, 0.2%si, 0.0%st I'm not sure why you want multiple CPUs in a virtual environment when you can clone a new one for each task. I'm not sure how to do that. Do you have a hint or page to refer me to? -- Bruce -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Kernel options help
To list, I built a new LFS 6.8 and everything is kosher save for some slowness. I built an x86_64 kernel (2.6.38.2) all on a Xen host (5.6.100) on a Quad Proc Xeon. It boots with no issues until I try to configure a package. As an example, if I run the ./configure for the openssh package, it takes around 5 min just to configure. Making with -j4 tends to go well but still not as fast as I think it could. It shouldn't be a Xen issue as my LFS build system was the LFS live CD. Configuring under the liveCD proceeded at a normal rate. I've tried a few different kernel options but none seem to make any difference. Any help would be appreciated. -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Kernel options help
On Thu, Apr 14, 2011 at 08:13:06AM -0600, Dave Hajoglou wrote: To list, I built a new LFS 6.8 and everything is kosher save for some slowness. I built an x86_64 kernel (2.6.38.2) all on a Xen host (5.6.100) on a Quad Proc Xeon. It boots with no issues until I try to configure a package. As an example, if I run the ./configure for the openssh package, it takes around 5 min just to configure. Making with -j4 tends to go well but still not as fast as I think it could. It shouldn't be a Xen issue as my LFS build system was the LFS live CD. Configuring under the liveCD proceeded at a normal rate. I've tried a few different kernel options but none seem to make any difference. Any help would be appreciated. I wouldn't expect the kernel config to make a lot of difference. The whole system is 64-bit ? It rather sounds like you might have masses of memory on a 32-bit system, and all the time is taken up in bounce-buffers (to address more than 4GB). 'top' might help - while configure is running, to see what where the cycles are being spent - a lot of time in wait might imply a disk or filesystem problem. Or, perhaps there is a lot of background activity. Maybe it's a question of configuring xen differently, e.g. pinning a cpu core (I see that mentioned in the Arch wiki, but I've no idea how to do it). ĸen -- das eine Mal als Tragödie, das andere Mal als Farce -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Kernel options help
Dave Hajoglou wrote: On Thu, Apr 14, 2011 at 8:13 AM, Dave Hajoglou dhajog...@gmail.com wrote: To list, I built a new LFS 6.8 and everything is kosher save for some slowness. I built an x86_64 kernel (2.6.38.2) all on a Xen host (5.6.100) on a Quad Proc Xeon. It boots with no issues until I try to configure a package. As an example, if I run the ./configure for the openssh package, it takes around 5 min just to configure. Looks more like around 34 min to configure. # time ./configure...: real34m19.631s user0m22.352s sys 36m1.760s # time make -j4 real1m27.764s user0m37.722s sys 5m3.909s There is definitely something wrong. On a production LFS system running in a virtual envronment, I get: real0m18.514s user0m8.984s sys 0m2.697s cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 6 model name : QEMU Virtual CPU version 0.9.1 stepping: 3 cpu MHz : 2260.701 cache size : 32 KB fdiv_bug: no hlt_bug : no f00f_bug: no coma_bug: no fpu : yes fpu_exception : yes cpuid level : 4 wp : yes flags : fpu de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pse36 clflush mmx fxsr sse sse2 syscall nx lm pni hypervisor bogomips: 4521.40 clflush size: 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management: I'm not sure why you want multiple CPUs in a virtual environment when you can clone a new one for each task. -- Bruce -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page