Invalid memory stats from vmstat and sysctl vm.vmtotal?
We're seeing some impossible memory usage stats reported on machines here from vmstat and sysctl vm.vmtotal. We have machines reporting to be using 31GB total when they only have 8GB physical and are not using any swap. Here's an output from one of our machines:- vmstat -c 2 -w 1 -n 0 procs memory page faults cpu r b w avmfre flt re pi pofr sr in sy cs us sy id 0 0 0 31768M 2112M 586 0 0 0 421 0 106 270 569 0 6 94 0 0 0 31768M 2112M 2 0 0 0 0 0 370 8139 3996 0 1 99 The raw output is:- vmstat -c 2 -w 1 -n 0 -H procs memory page faults cpu r b w avmfre flt re pi pofr sr in sy cs us sy id 0 0 0 32530228 2162524 586 0 0 0 421 0 106 270 569 0 6 94 0 0 0 32530228 2162524 2 0 0 0 0 0 286 8234 4347 0 1 99 Top shows:- last pid: 6665; load averages: 0.00, 0.00, 0.01 up 80+01:24:12 09:35:28 1893 processes:1 running, 1892 sleeping CPU: 0.0% user, 0.0% nice, 0.3% system, 0.0% interrupt, 99.7% idle Mem: 3754M Active, 84M Inact, 1976M Wired, 4K Cache, 2109M Free Swap: 4096M Total, 4096M Free sysctl vm.vmtotal vm.vmtotal: System wide totals computed every five seconds: (values in kilobytes) === Processes: (RUNQ: 1 Disk Wait: 0 Page Wait: 0 Sleep: 1893) Virtual Memory: (Total: 1106403532K Active: 32540260K) Real Memory:(Total: 4563648K Active: 3921644K) Shared Virtual Memory: (Total: 19976K Active: 16396K) Shared Real Memory: (Total: 9040K Active: 8436K) Free Memory Pages: 2161740K As mentioned this machine has 8GB of ram and according to both top and swapinfo is using no swap at all From dmesg:- real memory = 8589934592 (8192 MB) avail memory = 823536 (7873 MB) swapinfo Device 1K-blocks UsedAvail Capacity /dev/gptid/09f211f7-39ce-11e0-8 41943040 4194304 0% uname -a FreeBSD test 8.2-RELEASE FreeBSD 8.2-RELEASE #2: Thu Mar 24 17:28:55 UTC 2011 root@test:/usr/obj/usr/src/sys/MULTIPLAY amd64 sysctl hw.pagesize hw.pagesize: 4096 It looks like it may be out by a factor of 4, possibly due to the fact the its a 4k page size not 1k as indicated by the vmstat man page:- memory Information about the usage of virtual and real memory. Virtual pages (reported in units of 1024 bytes) are considered active if they belong to processes which are running or have run in the last 20 seconds. avm active virtual pages fre size of the free list Totalling up RSS from ps axo rss gives a total in the region of that if the vm stats are out by a factor of 4, in this case it should be: 8132557 which is 7.75GB a much more realistic value. Am I totally missing something or is there problem here? Regards Steve This e.mail is private and confidential between Multiplay (UK) Ltd. and the person or entity to whom it is addressed. In the event of misdirection, the recipient is prohibited from using, copying, printing or otherwise disseminating it or any information contained in it. In the event of misdirection, illegible or incomplete transmission please telephone +44 845 868 1337 or return the E.mail to postmas...@multiplay.co.uk. ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to freebsd-hackers-unsubscr...@freebsd.org
Re: sysctl description spillover and also setting the sysctl ?
On Friday, November 25, 2011 2:36:30 am Jason Hellenthal wrote: Found a troubling result of the following and figured someone might want to take a look. Pay close attention to the output and behavior. sysctl net.inet.udp.blackhole=0 sysctl net.inet.udp.blackhole sysctl -d net.inet.udp.blackhole=1 sysctl net.inet.udp.blackhole Is this expected ? should it not just display the description instead of adjusting ? as well not display the description like it is adjusting the description too ? Hah, cute. It should probably fail with an error if you do something like that, yes. -- John Baldwin ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to freebsd-hackers-unsubscr...@freebsd.org
Re: sysctl description spillover and also setting the sysctl ?
On Wed, Nov 30, 2011 at 11:52:46AM -0500, John Baldwin wrote: On Friday, November 25, 2011 2:36:30 am Jason Hellenthal wrote: Found a troubling result of the following and figured someone might want to take a look. Pay close attention to the output and behavior. sysctl net.inet.udp.blackhole=0 sysctl net.inet.udp.blackhole sysctl -d net.inet.udp.blackhole=1 sysctl net.inet.udp.blackhole Is this expected ? should it not just display the description instead of adjusting ? as well not display the description like it is adjusting the description too ? Hah, cute. It should probably fail with an error if you do something like that, yes. Yeah thats what I thought about it to but the more I thought about it, if it just displayed the values changing instead of the description when =N is supplied I think that would be acceptable to. 0 - 1 in this case. Or possibly sys.oid: 0 - 1 # Description since sysctl.conf(5) also takes comments like that. Not really thats something at the top of the list for fixes though. Low fruit. Food for thought. ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to freebsd-hackers-unsubscr...@freebsd.org
[CFT] pkgng alpha2
Hi all, We are releasing pkgng (the next pkg_install) alpha2 to the world and we want you to test it! There is no good method to test it: use it as you would in the real world. Of course, you are encouraged to backup your data or test it in some kind of virtualized environment. After using it for some time, you will certainly find bugs. You can report them on the issues tracker [1]. If you find missing features, that is things you can't do with pkgng but can with pkg_install, you can also report them. New features are not the expected outcome of this call, as we want to release a final version ASAP. FYI, an alpha3 should follow shortly to fix issues in alpha3 and test additional features. After that, there will be a feature freeze with beta1. Getting started: You can download or git clone the source code of pkgng on the github page [2]. Then, a boring `make' followed by `make install' will do it. If you have some packages installed by pkg_add, you can convert the old database to the pkgng database with the 'pkg2ng' shell script in the ports/ folder. You can also add packages from the ports tree (with bsd.pkgng.mk) or with a pkgng repository. All is documented in the README and manpages. If you are a newcomer to pkgng, this doc reading step is also valuable to us. Indeed, if you fight to get the right infos, or if some things feel counter intuitive, we should improve it! Which brings me to the topic of contributing to pkgng. The best thing you can do is to write down the documentation you would have loved to read while testing pkgng! And of course, if you have a patch with your bug report, it is much appreciated. If you read this entire mail and wonder what is this pkgng thing, you can read the wiki page [3], bapt's presentation from BSDCan [4], EuroBSDCon [5] [6] and browse the source code. Regards, Julien, on behalf of the pkgng team. And remember, we _do_ want to hear back from you! Please also note that it is still alpha code and it can kill kitten and puppies. You are warned ;-) [1] : https://github.com/pkgng/pkgng/issues [2] : https://github.com/pkgng/pkgng [3] : http://wiki.freebsd.org/pkgng [4] : http://people.freebsd.org/~bapt/pkgng-bsdcan2011.pdf [5] : http://wiki.freebsd.org/201110DevSummit/Ports?action=AttachFiledo=gettarget=pkgng-devsummit.pdf [6] : http://wiki.freebsd.org/201110DevSummit?action=AttachFiledo=gettarget=pkgng-devsummit-track.pdf ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to freebsd-hackers-unsubscr...@freebsd.org
Re: sysctl description spillover and also setting the sysctl ?
On 11/30/11 7:48 PM, Jason Hellenthal wrote: On Wed, Nov 30, 2011 at 11:52:46AM -0500, John Baldwin wrote: On Friday, November 25, 2011 2:36:30 am Jason Hellenthal wrote: Found a troubling result of the following and figured someone might want to take a look. Pay close attention to the output and behavior. sysctl net.inet.udp.blackhole=0 sysctl net.inet.udp.blackhole sysctl -d net.inet.udp.blackhole=1 sysctl net.inet.udp.blackhole Is this expected ? should it not just display the description instead of adjusting ? as well not display the description like it is adjusting the description too ? Hah, cute. It should probably fail with an error if you do something like that, yes. Yeah thats what I thought about it to but the more I thought about it, if it just displayed the values changing instead of the description when =N is supplied I think that would be acceptable to. 0 - 1 in this case. Or possibly sys.oid: 0 - 1 # Description since sysctl.conf(5) also takes comments like that. Not really thats something at the top of the list for fixes though. Low fruit. Food for thought. Perhaps you would kindly provide a patch for this ? I might have a look at it if nobody does, sounds trivial enough that I might be able to handle it ;) ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to freebsd-hackers-unsubscr...@freebsd.org
Re: sysctl description spillover and also setting the sysctl ?
On 11/30/11 7:48 PM, Jason Hellenthal wrote:
On Wed, Nov 30, 2011 at 11:52:46AM -0500, John Baldwin wrote:
On Friday, November 25, 2011 2:36:30 am Jason Hellenthal wrote:
Found a troubling result of the following and figured someone might want to
take a look.
Pay close attention to the output and behavior.
sysctl net.inet.udp.blackhole=0
sysctl net.inet.udp.blackhole
sysctl -d net.inet.udp.blackhole=1
sysctl net.inet.udp.blackhole
Is this expected ? should it not just display the description instead of
adjusting ? as well not display the description like it is adjusting the
description too ?
Hah, cute. It should probably fail with an error if you do something like
that, yes.
Yeah thats what I thought about it to but the more I thought about it, if it
just displayed the values changing instead of the description when =N is
supplied I think that would be acceptable to. 0 - 1 in this case. Or
possibly sys.oid: 0 - 1 # Description since sysctl.conf(5) also takes
comments like that.
Not really thats something at the top of the list for fixes though. Low
fruit. Food for thought.
Ok so I'm totally not a dev, don't take my findings for granted, but:
(gdb) run
The program being debugged has been started already.
Start it from the beginning? (y or n) y
Starting program: /usr/src/sbin/sysctl/sysctl -d net.inet.udp.blackhole=1
Breakpoint 5, show_var (oid=0x7fffe1d0, nlen=4) at
/usr/src/sbin/sysctl/sysctl.c:549
549 {
(gdb) s
562 bzero(buf, BUFSIZ);
(gdb) s
563 bzero(name, BUFSIZ);
(gdb) s
564 qoid[0] = 0;
(gdb) s
565 memcpy(qoid + 2, oid, nlen * sizeof(int));
(gdb) s
567 qoid[1] = 1;
(gdb) s
568 j = sizeof(name);
(gdb) s
569 i = sysctl(qoid, nlen + 2, name, j, 0, 0);
(gdb) s
570 if (i || !j)
(gdb) s
573 if (Nflag) {
(gdb) s
578 if (eflag)
(gdb) s
581 sep = : ;
(gdb) s
583 if (dflag) {/* just print description */
(gdb) s
584 qoid[1] = 5;
(gdb) s
585 j = sizeof(buf);
(gdb) s
586 i = sysctl(qoid, nlen + 2, buf, j, 0, 0);
(gdb) s
587 if (!nflag)
(gdb) s
588 printf(%s%s, name, sep);
(gdb) s
net.inet.udp.blackhole:
589 printf(%s, buf);
(gdb) s
Do not send port unreachables for refused connects
590 return (0);
(gdb) s
719 }
At this point, we have shown the description and should exit.
Instead, we're calling parse() again ?
(gdb) s
parse (string=0x7fffee07 net.inet.udp.blackhole) at
/usr/src/sbin/sysctl/sysctl.c:299
299 if (sysctl(mib, len, 0, 0, newval, newsize) == -1) {
(gdb) s
318 if (!bflag)
(gdb) s
319 printf( - );
(gdb) s
- 320 i = nflag;
(gdb) s
321 nflag = 1;
(gdb) s
322 j = show_var(mib, len);
(gdb) s
Breakpoint 5, show_var (oid=0x7fffe1d0, nlen=4) at
/usr/src/sbin/sysctl/sysctl.c:549
549 {
(gdb) s
562 bzero(buf, BUFSIZ);
(gdb) s
563 bzero(name, BUFSIZ);
(gdb) s
564 qoid[0] = 0;
(gdb) s
565 memcpy(qoid + 2, oid, nlen * sizeof(int));
(gdb) s
567 qoid[1] = 1;
(gdb) s
568 j = sizeof(name);
(gdb) s
569 i = sysctl(qoid, nlen + 2, name, j, 0, 0);
(gdb) s
570 if (i || !j)
(gdb) s
573 if (Nflag) {
(gdb) s
578 if (eflag)
(gdb) s
581 sep = : ;
(gdb) s
583 if (dflag) {/* just print description */
(gdb) s
584 qoid[1] = 5;
(gdb) s
585 j = sizeof(buf);
(gdb) s
586 i = sysctl(qoid, nlen + 2, buf, j, 0, 0);
(gdb) s
587 if (!nflag)
(gdb) s
589 printf(%s, buf);
(gdb) s
Do not send port unreachables for refused connects
590 return (0);
(gdb) s
719 }
(gdb) s
parse (string=0x7fffee07 net.inet.udp.blackhole) at
/usr/src/sbin/sysctl/sysctl.c:323
323 if (!j !bflag)
(gdb) s
324 putchar('\n');
(gdb) s
__sputc (_c=10, _p=0x80085a810) at stdio.h:457
457 if (--_p-_w = 0 || (_p-_w = _p-_lbfsize (char)_c
!= '\n'))
(gdb) s
460 return (__swbuf(_c, _p));
(gdb) s
461 }
(gdb) s
parse (string=0x7fffee07 net.inet.udp.blackhole) at
/usr/src/sbin/sysctl/sysctl.c:325
325 nflag = i;
(gdb) s
327 }
(gdb) s
main (argc=0, argv=0x7fffeb48) at /usr/src/sbin/sysctl/sysctl.c:154
154 while (argc-- 0)
(gdb) s
156 exit(warncount);
(gdb) s
Program exited normally.
I think that the issue here is that sysctl continues being run after
displaying the description at line 588.
I would assume that upon reaching the closing hyphen at line 719 ,
Re: Invalid memory stats from vmstat and sysctl vm.vmtotal?
On Wed, Nov 30, 2011 at 12:39:10PM -, Steven Hartland wrote: We're seeing some impossible memory usage stats reported on machines here from vmstat and sysctl vm.vmtotal. We have machines reporting to be using 31GB total when they only have 8GB physical and are not using any swap. Here's an output from one of our machines:- vmstat -c 2 -w 1 -n 0 procs memory page faults cpu r b w avmfre flt re pi pofr sr in sy cs us sy id 0 0 0 31768M 2112M 586 0 0 0 421 0 106 270 569 0 6 94 0 0 0 31768M 2112M 2 0 0 0 0 0 370 8139 3996 0 1 99 The raw output is:- vmstat -c 2 -w 1 -n 0 -H procs memory page faults cpu r b w avmfre flt re pi pofr sr in sy cs us sy id 0 0 0 32530228 2162524 586 0 0 0 421 0 106 270 569 0 6 94 0 0 0 32530228 2162524 2 0 0 0 0 0 286 8234 4347 0 1 99 Top shows:- last pid: 6665; load averages: 0.00, 0.00, 0.01 up 80+01:24:12 09:35:28 1893 processes:1 running, 1892 sleeping CPU: 0.0% user, 0.0% nice, 0.3% system, 0.0% interrupt, 99.7% idle Mem: 3754M Active, 84M Inact, 1976M Wired, 4K Cache, 2109M Free Swap: 4096M Total, 4096M Free sysctl vm.vmtotal vm.vmtotal: System wide totals computed every five seconds: (values in kilobytes) === Processes: (RUNQ: 1 Disk Wait: 0 Page Wait: 0 Sleep: 1893) Virtual Memory: (Total: 1106403532K Active: 32540260K) Real Memory:(Total: 4563648K Active: 3921644K) Shared Virtual Memory: (Total: 19976K Active: 16396K) Shared Real Memory: (Total: 9040K Active: 8436K) Free Memory Pages: 2161740K As mentioned this machine has 8GB of ram and according to both top and swapinfo is using no swap at all From dmesg:- real memory = 8589934592 (8192 MB) avail memory = 823536 (7873 MB) swapinfo Device 1K-blocks UsedAvail Capacity /dev/gptid/09f211f7-39ce-11e0-8 41943040 4194304 0% uname -a FreeBSD test 8.2-RELEASE FreeBSD 8.2-RELEASE #2: Thu Mar 24 17:28:55 UTC 2011 root@test:/usr/obj/usr/src/sys/MULTIPLAY amd64 sysctl hw.pagesize hw.pagesize: 4096 It looks like it may be out by a factor of 4, possibly due to the fact the its a 4k page size not 1k as indicated by the vmstat man page:- memory Information about the usage of virtual and real memory. Virtual pages (reported in units of 1024 bytes) are considered active if they belong to processes which are running or have run in the last 20 seconds. avm active virtual pages fre size of the free list Totalling up RSS from ps axo rss gives a total in the region of that if the vm stats are out by a factor of 4, in this case it should be: 8132557 which is 7.75GB a much more realistic value. Am I totally missing something or is there problem here? Regards Steve This goes along with the thoughts I had about 4 months ago tending to some zfs statistics as well top showing greater than 100% actual CPU usage. This is a big pet peave of mine. Its like saying you ate 134% of a bannanna when in all reallity it is impossible. You can never have more than 100% usage of anything and when seen is a clear notice that some math is considerably incorrect leading to other such miscalculations to be performed. Things like the above already have checks in place that ensure no boundries are being crossed/overflowed or underrun but it surely makes processing results building future products a bitch. One instance is the calculation of threads for example firefox can be seen using upto or more 338% of the CPU. Thats impossible its like saying anyones CPU grew by 400%. ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to freebsd-hackers-unsubscr...@freebsd.org
