Re: Authentication tried for XXX with correct key but not from a permitted host
On 11/07/2010 04:04:57, Dan Langille wrote: That asked, I know if I move the key to the top of the ~/.ssh/authorized_keys file, the message is no longer logged. Further investigation reveals that if a line of the form: from=10..etc appears before the key being used to log in, the message will appear. Usually the from='10.0.0.100' tag should be inserted at the beginning of the line for each key it should affect. It shouldn't do anything on a line on its own -- in fact that should be a syntax error. The behaviour you're seeing sounds like something new: it isn't what sshd(8) describes in the section on AUTHORIZED_KEYS FILE FORMAT. This new behaviour sounds as if it could be quite useful for easing the management of complicated authorised_keys files, but I'd have expected some sort of notice somewhere. I can't see anything relevant in the release notes for OpenSSH for versions 5.0, 5.1, 5.3, 5.3, 5.4 or 5.5 [Eg. http://www.openssh.org/txt/release-5.4 -- 8.1-PRERELEASE has OpenSSH 5.4p1 bundled]. Nor anything in any of the ssh(1), ssh_config(1), sshd(8), sshd_config(8) man pages. Maybe it's a bug, but one that has fortuitously useful effects. Cheers, Mathew -- Dr Matthew J Seaman MA, D.Phil. 7 Priory Courtyard Flat 3 PGP: http://www.infracaninophile.co.uk/pgpkey Ramsgate JID: matt...@infracaninophile.co.uk Kent, CT11 9PW signature.asc Description: OpenPGP digital signature
dmg handling
Howdy! I got a file made on mac, sized at 142 mb and with dmg extension. It is a dump of 4 gb SD card with firmware upgrade files. File says: disk1s1.dmg: VAX COFF executable not stripped So, compressed. The file system on that original card is ext3. Linux people use dmg2img utility to decompress and then mount it with options like: mount -o loop -t hfsplus image dir If correct, I could try to bunzip2 it first, then rsync to sheevaplug node (ubuntu on it) and mount finally. The final step would be to write that ima- ge to new SD card and use it to upgrade the gadget. I was sure that simple dd would suffice, but more I read, more I see it's wrong. I also assume that I cannot write that image to smaller, 2 gb SD card. Any idea what to do in this situation? Best regards Zoran ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
Serious zfs slowdown when mixed with another file system (ufs/msdosfs/etc.).
This is on clean FreeBSD 8.1 RC2, amd64, with 4GB memory. The closest I found by Googling was this: http://forums.freebsd.org/showthread.php?t=9935 And it talks about all kinds of little tweaks, but in the end, the only thing that actually works is the stupid 1-line perl code that forces the kernal to free the memory allocated to (non-zfs) disk cache, which is the Inactive memory in top. I have a 4-disk raidz pool, but that's unlikely to matter. Try to copy large files from non-zfs disk to zfs disk. FreeBSD will cache the data read from non-zfs disk in memory, and free memory will go down. This is as expected, obviously. Once there's very little free memory, one would expect whatever is more important to kick out the cached data (Inact) and make memory available. But when almost all of the memory is taken by disk cache (of non-zfs file system), ZFS disks start threshing like mad and the write throughput goes down in 1-digit MB/second. I believe it should be extremely easy to duplicate. Just plug in a big USB drive formatted in UFS (msdosfs will likely do the same), and copy large files from that USB drive to zfs pool. Right after clean boot, gstat will show something like 20+MB/s movement from USB device (da*), and occasional bursts of activity on zpool devices at very high rate. Once free memory is exhausted, zpool devices will change to constant low-speed activity, with disks threshing about constantly. I tried enabling/disabling prefetch, messing with vnode counts, zfs.vdev.min/max_pending, etc. The only thing that works is that stupid perl 1-liner (perl -e '$x=xx15'), which returns the activity to that seen right after a clean boot. It doesn't last very long, though, as the disk cache again consumes all the memory. Copying files between zfs devices doesn't seem to affect anything. I understand zfs subsystem has its own memory/cache management. Can a zfs expert please comment on this? And is there a way to force the kernel to not cache non-zfs disk data? --rich ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
syslogs altlog_proglist and isc-dhcpd logging for FreeBSD
Hello, since isc-dhcpd-4.1.1 promised ipv6, I wanted to replace my existing DHCP servers with this new version. I'm running chrooted. My problem was with logging. dhcpd is very noisy and setting log-facility local1 in dhcpd.conf doesn't work out of the box (*) because of the chrooted environment. But some good guys already coded everything which is needed to have dhcpd logging with chrooted. - syslogd has the -l switch which enables to place an additional log socket into the chrooted enivronment. - /etc/rc.d/syslogd already knows about this and has the variable altlog_proglist, which checks for possible chrooted daemons The problems are: - /etc/rc.d/syslogd has the altlog_proglist hard coded - /etc/rc.d/syslogd checks for daemons in rc.conf which have flags any_chrootdir, but rc.d/isc-dhcpd uses dhcpd_rootdir. So here's the view simple lines that make dhcpd logging working with individula log-facility configs: --- etc/rc.d/syslogd2009-09-06 02:47:31.0 +0200 +++ etc/rc.d/syslogd2010-07-11 21:27:46.477366986 +0200 @@ -1,6 +1,6 @@ #!/bin/sh # -# $FreeBSD: src/etc/rc.d/syslogd,v 1.13.2.1 2009/08/03 08:13:06 kensmith Exp $ +# $FreeBSD: src/etc/rc.d/syslogd,v 1.13.2.1.4.1 2010/06/14 02:09:06 kensmith Exp $ # # PROVIDE: syslogd @@ -19,7 +19,9 @@ sockfile=/var/run/syslogd.sockets evalargs=rc_flags=\\`set_socketlist\` \$rc_flags\ -altlog_proglist=named + +load_rc_config $name +altlog_proglist=${syslogd_altlog_proglist:-named} syslogd_precmd() { --- etc/defaults/rc.conf2009-11-01 15:08:40.0 +0100 +++ etc/defaults/rc.conf2010-07-11 21:30:04.373974162 +0200 @@ -255,6 +255,7 @@ syslogd_enable=YES # Run syslog daemon (or NO). syslogd_program=/usr/sbin/syslogd # path to syslogd, if you want a different one. syslogd_flags=-s # Flags to syslogd (if enabled). +syslogd_altlog_proglist=named # Check vor chrooted daemons and place additional socket inetd_enable=NO# Run the network daemon dispatcher (YES/NO). inetd_program=/usr/sbin/inetd # path to inetd, if you want a different one. inetd_flags=-wW -C 60 # Optional flags to inetd --- etc/rc.d/isc-dhcpd.orig 2010-07-08 13:03:45.0 +0200 +++ etc/rc.d/isc-dhcpd 2010-07-11 20:41:36.0 +0200 @@ -32,7 +32,7 @@ dhcpd_chroot_enable=${dhcpd_chroot_enable:-NO} # runs chrooted? dhcpd_devfs_enable=${dhcpd_devfs_enable:-YES} # devfs if available? -dhcpd_rootdir=${dhcpd_rootdir:-/var/db/${name}}# directory to run in +dhcpd_rootdir=${dhcpd_chrootdir:-/var/db/${name}} # directory to run in # dhcpd_includedir=# directory for included config files safe_run ()# rc command [args...] Is it possible to get these changes into base system? @wxs Any objections changing dhacpd_rootdir into dhcpd_chrootdir variable? Shall I file a PR? Thanks, -Harry P.S.: For the records, here another possibility to make dhcpd use different syslog facility in chrooted environmen: (*) Chaging the syslog facility of dhcpd with log-facility local7; in dhcpd.conf doesn't work for chrooted dhcpd. At startup, it uses the local datagram syslogd socket /dev/log (/var/run/syslog.sockets). The syslog facility change is done after changeroot took place, so in the chrooted environment there is no syslogd reachable. To change the default syslog facility from LOG_DAEMON to LOG_LOCAL7 add the following to the ports Makefile: CONFIGURE_ENV= CPPFLAGS=-DDHCPD_LOG_FACILITY=LOG_LOCAL7 .. *snip* signature.asc Description: OpenPGP digital signature
Re: 8.x grudges
In message: aanlktikdj39liaffibdwkfa1vgt4w7m8toxevjykh...@mail.gmail.com Garrett Cooper yanef...@gmail.com writes: : On Wed, Jul 7, 2010 at 1:17 PM, Mikhail T. mi+t...@aldan.algebra.com wrote: : 07.07.2010 14:59, Jeremy Chadwick ???(??): : : FREEBSD_COMPAT7 kernel option is, apparently, a requirement (and : thus not an option) -- the kernel-config files, that worked with : 7.x, break without this option in them (in addition to all the : nuisance, that's documented in UPDATING -- which, somehow, makes : the breakage acceptable). config(8) would not warn about this, but : kernel build fails. : : : We don't use this option (meaning it's removed from our kernels). It's : definitely not required. All it does is ensure your kernel can : comprehend executables/binaries built on 7.x. : : : Attached is the kernel config-file (i386), that worked fine under 7.x. The : kernel-compile will break (some *freebsd7* structs undefined), without the : COMPAT_FREEBSD7 option. Try it for yourself... : : options SYSVSHM # SYSV-style shared memory : options SYSVMSG # SYSV-style message queues : options SYSVSEM # SYSV-style semaphores : : Those require COMPAT_FREEBSD7. This does seem like a bug: : : static struct syscall_helper_data shm_syscalls[] = { : SYSCALL_INIT_HELPER(shmat), : SYSCALL_INIT_HELPER(shmctl), : SYSCALL_INIT_HELPER(shmdt), : SYSCALL_INIT_HELPER(shmget), : #if defined(COMPAT_FREEBSD4) || defined(COMPAT_FREEBSD5) || \ : defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD7) : SYSCALL_INIT_HELPER(freebsd7_shmctl), : #endif : : The check should be for COMPAT_FREEBSD7 only I would think. : : Apart from that, everything else should work without it I would think. You would think that, but you'd be wrong. In general, if you have COMPAT_FREEBSDx defined, you need all COMPAT_FREEBSDy for y x defined. The reason for this is that we name the compat shim for the version where it was removed, but it is needed for all prior versions. freebsd7_shmctl is needed to emulate the earlier versions as well... This is why we'd like to move to something more like COMPAT_MIN_FREEBSD=z, but there's hooks into the config system and syscall tables that make it tricky... Warner ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
Re: Serious zfs slowdown when mixed with another file system (ufs/msdosfs/etc.).
On Sun, Jul 11, 2010 at 11:25:12AM -0700, Richard Lee wrote: This is on clean FreeBSD 8.1 RC2, amd64, with 4GB memory. The closest I found by Googling was this: http://forums.freebsd.org/showthread.php?t=9935 And it talks about all kinds of little tweaks, but in the end, the only thing that actually works is the stupid 1-line perl code that forces the kernal to free the memory allocated to (non-zfs) disk cache, which is the Inactive memory in top. I have a 4-disk raidz pool, but that's unlikely to matter. Try to copy large files from non-zfs disk to zfs disk. FreeBSD will cache the data read from non-zfs disk in memory, and free memory will go down. This is as expected, obviously. Once there's very little free memory, one would expect whatever is more important to kick out the cached data (Inact) and make memory available. But when almost all of the memory is taken by disk cache (of non-zfs file system), ZFS disks start threshing like mad and the write throughput goes down in 1-digit MB/second. I believe it should be extremely easy to duplicate. Just plug in a big USB drive formatted in UFS (msdosfs will likely do the same), and copy large files from that USB drive to zfs pool. Right after clean boot, gstat will show something like 20+MB/s movement from USB device (da*), and occasional bursts of activity on zpool devices at very high rate. Once free memory is exhausted, zpool devices will change to constant low-speed activity, with disks threshing about constantly. I tried enabling/disabling prefetch, messing with vnode counts, zfs.vdev.min/max_pending, etc. The only thing that works is that stupid perl 1-liner (perl -e '$x=xx15'), which returns the activity to that seen right after a clean boot. It doesn't last very long, though, as the disk cache again consumes all the memory. Copying files between zfs devices doesn't seem to affect anything. I understand zfs subsystem has its own memory/cache management. Can a zfs expert please comment on this? And is there a way to force the kernel to not cache non-zfs disk data? I believe you may be describing two separate issues: 1) ZFS using a lot of memory but not freeing it as you expect 2) Lack of disk I/O scheduler For (1), try this in /boot/loader.conf and reboot: # Disable UMA (uma(9)) for ZFS; amd64 was moved to exclusively use UMA # on 2010/05/24. # http://lists.freebsd.org/pipermail/freebsd-stable/2010-June/057162.html vfs.zfs.zio.use_uma=0 For (2), may try gsched_rr: http://svnweb.freebsd.org/viewvc/base/releng/8.1/sys/geom/sched/README?view=markup -- | Jeremy Chadwick j...@parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, USA | | Making life hard for others since 1977. PGP: 4BD6C0CB | ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
RE: Serious zfs slowdown when mixed with another file system (ufs/msdosfs/etc.).
Using Jeremy's suggestion as follows: 1) ZFS using a lot of memory but not freeing it as you expect For (1), try this in /boot/loader.conf and reboot: vfs.zfs.zio.use_uma=0 ... works like a charm for me. Thank you. Scott -Original Message- From: owner-freebsd-sta...@freebsd.org [mailto:owner-freebsd-sta...@freebsd.org] On Behalf Of Jeremy Chadwick Sent: Sunday, July 11, 2010 1:48 PM To: Richard Lee Cc: freebsd-stable@freebsd.org Subject: Re: Serious zfs slowdown when mixed with another file system (ufs/msdosfs/etc.). On Sun, Jul 11, 2010 at 11:25:12AM -0700, Richard Lee wrote: This is on clean FreeBSD 8.1 RC2, amd64, with 4GB memory. The closest I found by Googling was this: http://forums.freebsd.org/showthread.php?t=9935 And it talks about all kinds of little tweaks, but in the end, the only thing that actually works is the stupid 1-line perl code that forces the kernal to free the memory allocated to (non-zfs) disk cache, which is the Inactive memory in top. I have a 4-disk raidz pool, but that's unlikely to matter. Try to copy large files from non-zfs disk to zfs disk. FreeBSD will cache the data read from non-zfs disk in memory, and free memory will go down. This is as expected, obviously. Once there's very little free memory, one would expect whatever is more important to kick out the cached data (Inact) and make memory available. But when almost all of the memory is taken by disk cache (of non-zfs file system), ZFS disks start threshing like mad and the write throughput goes down in 1-digit MB/second. I believe it should be extremely easy to duplicate. Just plug in a big USB drive formatted in UFS (msdosfs will likely do the same), and copy large files from that USB drive to zfs pool. Right after clean boot, gstat will show something like 20+MB/s movement from USB device (da*), and occasional bursts of activity on zpool devices at very high rate. Once free memory is exhausted, zpool devices will change to constant low-speed activity, with disks threshing about constantly. I tried enabling/disabling prefetch, messing with vnode counts, zfs.vdev.min/max_pending, etc. The only thing that works is that stupid perl 1-liner (perl -e '$x=xx15'), which returns the activity to that seen right after a clean boot. It doesn't last very long, though, as the disk cache again consumes all the memory. Copying files between zfs devices doesn't seem to affect anything. I understand zfs subsystem has its own memory/cache management. Can a zfs expert please comment on this? And is there a way to force the kernel to not cache non-zfs disk data? I believe you may be describing two separate issues: 1) ZFS using a lot of memory but not freeing it as you expect 2) Lack of disk I/O scheduler For (1), try this in /boot/loader.conf and reboot: # Disable UMA (uma(9)) for ZFS; amd64 was moved to exclusively use UMA # on 2010/05/24. # http://lists.freebsd.org/pipermail/freebsd-stable/2010-June/057162.html vfs.zfs.zio.use_uma=0 For (2), may try gsched_rr: http://svnweb.freebsd.org/viewvc/base/releng/8.1/sys/geom/sched/README?view= markup -- | Jeremy Chadwick j...@parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, USA | | Making life hard for others since 1977. PGP: 4BD6C0CB | ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
More buzzing fun with snd_emu10kx - but now with more determinism!
Getting back to the thread I brought up before (with my now dead email address): http://unix.derkeiler.com/Mailing-Lists/FreeBSD/stable/2010-06/msg00036.html , I now have a more deterministic testcase for this issue. The problem appears to be with vchan-related code. If I start up 4+ applications on my machine that access the audio device, all goes wonky on the 4+ allocation (I was stress testing the nvidia driver to see whether or not it'd break with multiple instances of vlc, and stumbled on this by accident). So pushing the number of consumers of the audio subsystem forces a breakdown somewhere (even though the number of available hardware vchans is set to 16). I'll continue to look into this further as time permits. Thanks, -Garrett ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
Re: Serious zfs slowdown when mixed with another file system (ufs/msdosfs/etc.).
On Sun, Jul 11, 2010 at 01:47:57PM -0700, Jeremy Chadwick wrote: On Sun, Jul 11, 2010 at 11:25:12AM -0700, Richard Lee wrote: This is on clean FreeBSD 8.1 RC2, amd64, with 4GB memory. The closest I found by Googling was this: http://forums.freebsd.org/showthread.php?t=9935 And it talks about all kinds of little tweaks, but in the end, the only thing that actually works is the stupid 1-line perl code that forces the kernal to free the memory allocated to (non-zfs) disk cache, which is the Inactive memory in top. I have a 4-disk raidz pool, but that's unlikely to matter. Try to copy large files from non-zfs disk to zfs disk. FreeBSD will cache the data read from non-zfs disk in memory, and free memory will go down. This is as expected, obviously. Once there's very little free memory, one would expect whatever is more important to kick out the cached data (Inact) and make memory available. But when almost all of the memory is taken by disk cache (of non-zfs file system), ZFS disks start threshing like mad and the write throughput goes down in 1-digit MB/second. I believe it should be extremely easy to duplicate. Just plug in a big USB drive formatted in UFS (msdosfs will likely do the same), and copy large files from that USB drive to zfs pool. Right after clean boot, gstat will show something like 20+MB/s movement from USB device (da*), and occasional bursts of activity on zpool devices at very high rate. Once free memory is exhausted, zpool devices will change to constant low-speed activity, with disks threshing about constantly. I tried enabling/disabling prefetch, messing with vnode counts, zfs.vdev.min/max_pending, etc. The only thing that works is that stupid perl 1-liner (perl -e '$x=xx15'), which returns the activity to that seen right after a clean boot. It doesn't last very long, though, as the disk cache again consumes all the memory. Copying files between zfs devices doesn't seem to affect anything. I understand zfs subsystem has its own memory/cache management. Can a zfs expert please comment on this? And is there a way to force the kernel to not cache non-zfs disk data? I believe you may be describing two separate issues: 1) ZFS using a lot of memory but not freeing it as you expect 2) Lack of disk I/O scheduler For (1), try this in /boot/loader.conf and reboot: # Disable UMA (uma(9)) for ZFS; amd64 was moved to exclusively use UMA # on 2010/05/24. # http://lists.freebsd.org/pipermail/freebsd-stable/2010-June/057162.html vfs.zfs.zio.use_uma=0 For (2), may try gsched_rr: http://svnweb.freebsd.org/viewvc/base/releng/8.1/sys/geom/sched/README?view=markup -- | Jeremy Chadwick j...@parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, USA | | Making life hard for others since 1977. PGP: 4BD6C0CB | vfs.zfs.zio.use_uma is already 0. It looks to be the default, as I never touched it. And in my case, Wired memory is stable at around 1GB. It's the Inact memory that takes off, but only if reading from non-zfs file system. Without other file systems, I can keep moving files around and see no adverse slowdown. I can also scp huge files from another system into the zfs machine, and it doesn't affect memory usage (as reported by top), nor does it affect performance. As for gsched_rr, I don't believe this is related. There is only ONE access to the zfs devices (4 sata drives), which is purely a sequential write. The external USB HDD (UFS2) is a completely different device, and is doing purely sequential read. There is only one cp process doing anything at all. The FreeBSD system files aren't on either of these devices, either not that it's doing anything with the disk during this time. --rich ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
Re: Serious zfs slowdown when mixed with another file system (ufs/msdosfs/etc.).
On Sun, Jul 11, 2010 at 02:12:13PM -0700, Richard Lee wrote: On Sun, Jul 11, 2010 at 01:47:57PM -0700, Jeremy Chadwick wrote: On Sun, Jul 11, 2010 at 11:25:12AM -0700, Richard Lee wrote: This is on clean FreeBSD 8.1 RC2, amd64, with 4GB memory. The closest I found by Googling was this: http://forums.freebsd.org/showthread.php?t=9935 And it talks about all kinds of little tweaks, but in the end, the only thing that actually works is the stupid 1-line perl code that forces the kernal to free the memory allocated to (non-zfs) disk cache, which is the Inactive memory in top. I have a 4-disk raidz pool, but that's unlikely to matter. Try to copy large files from non-zfs disk to zfs disk. FreeBSD will cache the data read from non-zfs disk in memory, and free memory will go down. This is as expected, obviously. Once there's very little free memory, one would expect whatever is more important to kick out the cached data (Inact) and make memory available. But when almost all of the memory is taken by disk cache (of non-zfs file system), ZFS disks start threshing like mad and the write throughput goes down in 1-digit MB/second. I believe it should be extremely easy to duplicate. Just plug in a big USB drive formatted in UFS (msdosfs will likely do the same), and copy large files from that USB drive to zfs pool. Right after clean boot, gstat will show something like 20+MB/s movement from USB device (da*), and occasional bursts of activity on zpool devices at very high rate. Once free memory is exhausted, zpool devices will change to constant low-speed activity, with disks threshing about constantly. I tried enabling/disabling prefetch, messing with vnode counts, zfs.vdev.min/max_pending, etc. The only thing that works is that stupid perl 1-liner (perl -e '$x=xx15'), which returns the activity to that seen right after a clean boot. It doesn't last very long, though, as the disk cache again consumes all the memory. Copying files between zfs devices doesn't seem to affect anything. I understand zfs subsystem has its own memory/cache management. Can a zfs expert please comment on this? And is there a way to force the kernel to not cache non-zfs disk data? I believe you may be describing two separate issues: 1) ZFS using a lot of memory but not freeing it as you expect 2) Lack of disk I/O scheduler For (1), try this in /boot/loader.conf and reboot: # Disable UMA (uma(9)) for ZFS; amd64 was moved to exclusively use UMA # on 2010/05/24. # http://lists.freebsd.org/pipermail/freebsd-stable/2010-June/057162.html vfs.zfs.zio.use_uma=0 For (2), may try gsched_rr: http://svnweb.freebsd.org/viewvc/base/releng/8.1/sys/geom/sched/README?view=markup -- | Jeremy Chadwick j...@parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, USA | | Making life hard for others since 1977. PGP: 4BD6C0CB | vfs.zfs.zio.use_uma is already 0. It looks to be the default, as I never touched it. Okay, just checking, because the default did change at one point, as the link in my /boot/loader.conf denotes. Here's further confirmation (same thread), the first confirming on i386, the second confirming on amd64: http://lists.freebsd.org/pipermail/freebsd-stable/2010-June/057168.html http://lists.freebsd.org/pipermail/freebsd-stable/2010-June/057239.html And in my case, Wired memory is stable at around 1GB. It's the Inact memory that takes off, but only if reading from non-zfs file system. Without other file systems, I can keep moving files around and see no adverse slowdown. I can also scp huge files from another system into the zfs machine, and it doesn't affect memory usage (as reported by top), nor does it affect performance. Let me get this straight: The system has ZFS enabled (kernel module loaded), with a 4-disk raidz1 pool defined and used in the past (Wired being @ 1GB, due to ARC). The same system also has UFS2 filesystems. The ZFS pool vdevs consist of their own dedicated disks, and the UFS2 filesystems also have their own disk (which appears to be USB-based). When any sort of read I/O is done on the UFS2 filesystems, Inact skyrockets, and as a result this impacts performance of ZFS. If this is correct: can you remove USB from the picture and confirm the problem still happens? This is the first I've heard of the UFS caching mechanism spiraling out of control. By the way, all the stupid perl 1-liner does is make a process with an extremely large SIZE, and RES will grow to match it (more or less). The intention is to cause the VM to force a swap-out + free of memory by stressing the VM. Using 'x15', you'll find something like this: PID USERNAME THR PRI
Re: Serious zfs slowdown when mixed with another file system (ufs/msdosfs/etc.).
On Sun, Jul 11, 2010 at 02:45:46PM -0700, Jeremy Chadwick wrote: On Sun, Jul 11, 2010 at 02:12:13PM -0700, Richard Lee wrote: On Sun, Jul 11, 2010 at 01:47:57PM -0700, Jeremy Chadwick wrote: On Sun, Jul 11, 2010 at 11:25:12AM -0700, Richard Lee wrote: This is on clean FreeBSD 8.1 RC2, amd64, with 4GB memory. The closest I found by Googling was this: http://forums.freebsd.org/showthread.php?t=9935 And it talks about all kinds of little tweaks, but in the end, the only thing that actually works is the stupid 1-line perl code that forces the kernal to free the memory allocated to (non-zfs) disk cache, which is the Inactive memory in top. I have a 4-disk raidz pool, but that's unlikely to matter. Try to copy large files from non-zfs disk to zfs disk. FreeBSD will cache the data read from non-zfs disk in memory, and free memory will go down. This is as expected, obviously. Once there's very little free memory, one would expect whatever is more important to kick out the cached data (Inact) and make memory available. But when almost all of the memory is taken by disk cache (of non-zfs file system), ZFS disks start threshing like mad and the write throughput goes down in 1-digit MB/second. I believe it should be extremely easy to duplicate. Just plug in a big USB drive formatted in UFS (msdosfs will likely do the same), and copy large files from that USB drive to zfs pool. Right after clean boot, gstat will show something like 20+MB/s movement from USB device (da*), and occasional bursts of activity on zpool devices at very high rate. Once free memory is exhausted, zpool devices will change to constant low-speed activity, with disks threshing about constantly. I tried enabling/disabling prefetch, messing with vnode counts, zfs.vdev.min/max_pending, etc. The only thing that works is that stupid perl 1-liner (perl -e '$x=xx15'), which returns the activity to that seen right after a clean boot. It doesn't last very long, though, as the disk cache again consumes all the memory. Copying files between zfs devices doesn't seem to affect anything. I understand zfs subsystem has its own memory/cache management. Can a zfs expert please comment on this? And is there a way to force the kernel to not cache non-zfs disk data? I believe you may be describing two separate issues: 1) ZFS using a lot of memory but not freeing it as you expect 2) Lack of disk I/O scheduler For (1), try this in /boot/loader.conf and reboot: # Disable UMA (uma(9)) for ZFS; amd64 was moved to exclusively use UMA # on 2010/05/24. # http://lists.freebsd.org/pipermail/freebsd-stable/2010-June/057162.html vfs.zfs.zio.use_uma=0 For (2), may try gsched_rr: http://svnweb.freebsd.org/viewvc/base/releng/8.1/sys/geom/sched/README?view=markup -- | Jeremy Chadwick j...@parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, USA | | Making life hard for others since 1977. PGP: 4BD6C0CB | vfs.zfs.zio.use_uma is already 0. It looks to be the default, as I never touched it. Okay, just checking, because the default did change at one point, as the link in my /boot/loader.conf denotes. Here's further confirmation (same thread), the first confirming on i386, the second confirming on amd64: http://lists.freebsd.org/pipermail/freebsd-stable/2010-June/057168.html http://lists.freebsd.org/pipermail/freebsd-stable/2010-June/057239.html And in my case, Wired memory is stable at around 1GB. It's the Inact memory that takes off, but only if reading from non-zfs file system. Without other file systems, I can keep moving files around and see no adverse slowdown. I can also scp huge files from another system into the zfs machine, and it doesn't affect memory usage (as reported by top), nor does it affect performance. Let me get this straight: The system has ZFS enabled (kernel module loaded), with a 4-disk raidz1 pool defined and used in the past (Wired being @ 1GB, due to ARC). The same system also has UFS2 filesystems. The ZFS pool vdevs consist of their own dedicated disks, and the UFS2 filesystems also have their own disk (which appears to be USB-based). Yes, correct. I have: ad4 (An old 200GB SATA UFS2 main system drive) ad8, ad10, ad12, ad14 (1TB SATA drives) part of raidz1 and nothing else da0 is an external USB disk (1TB), but I don't think it's related to USB. Status looks like this: state: ONLINE scrub: none requested config: NAMESTATE READ WRITE CKSUM uchuu ONLINE 0 0 0 raidz1ONLINE 0 0 0 ad8 ONLINE
Re: Serious zfs slowdown when mixed with another file system (ufs/msdosfs/etc.).
Search the archives for the -stable, -current, and -fs mailing lists from the past 3 months. There are patches floating around to fix this. The ZFS code that monitors memory pressure currently only monitors the free amount, and completely ignores the inact and other not actually in use amounts. -- Freddie Cash fjwc...@gmail.com ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
Re: More buzzing fun with snd_emu10kx - but now with more determinism!
On 07/11/2010 17:03, Garrett Cooper wrote: The problem appears to be with vchan-related code. If I start up 4+ applications on my machine that access the audio device, all goes wonky on the 4+ allocation I can confirm this behaviour, which seems odd with hw.snd.maxautovchans defaulting to 16. It does not appear to be affected by increasing dev.pcm.0.play.vchans up fron the default of 2 (as I apparently did at some point up to 7.x), though reading sound(4) it's clear I don't fully understand vchans. A problem I encountered with snd_emu10kx in a clean 8.1RC2 install which was not present in any previous version is a faint rapid mechanical clicking sound adjustable with the cd mixer setting. The only non-default audio setting I have is in loader.conf: hint.emu10kx.0.multichannel_disabled=1 hint.emu10kx.1.disabled=1 pcm0: EMU10Kx DSP front PCM interface on emu10kx0 pcm0: TriTech TR28602 AC97 Codec (id = 0x54524123) pcm0: Codec features 5 bit master volume, no 3D Stereo Enhancement FreeBSD Audio Driver (newpcm: 32bit 2009061500/i386) Installed devices: pcm0: EMU10Kx DSP front PCM interface on emu10kx0 (4p:2v/1r:1v) default snddev flags=0x2e2AUTOVCHAN,BUSY,MPSAFE,REGISTERED,VPC [pcm0:play:dsp0.p0]: spd 48000, fmt 0x00200010, flags 0x2100, 0x0004 interrupts 726, underruns 0, feed 5, ready 0 [b:4096/2048/2|bs:4096/2048/2] channel flags=0x2100BUSY,HAS_VCHAN ... -- -Andrew J. Caines- Unix Systems Engineer a.j.cai...@halplant.com FreeBSD/Linux/Solaris, Web/Mail/Proxy/... http://halplant.com:2001/ Machines take me by surprise with great frequency - Alan Turing ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org