ataraid issue was Re: [netbsd-7] Critical issue with ffs+log
BERTRAND Joël a écrit : Thor Lancelot Simon a écrit : On Wed, Jan 20, 2016 at 05:15:09PM +0100, BERTRAND Jo??l wrote: Thor Lancelot Simon a ?crit : Ugh, ataraid. If you skip the first 1MB or thereabouts, do the drives have exactly the same contents? I suppose you say that ataraid is not usable in production... I will try to convert ataraid in raidframe configuration as soon as possible. I am not advocating that, I'm trying to gather information. Do the drives actually have the same contents at this point in time, or not? I don't know as this server if far away and doesn't reboot anymore. Kernel doesn't find /sbin/init. I will try to reboot this server next friday (CET). And after, I will check if both drives have same contents. Regards, JKB Hello, I cannot verify if both disk of my ataraid have same contents. But I have done some tests on two new NetBSD 7 installations with ataraid (same hardware, same disks). I have desynchronized raid on the second of these workstations (by entering in kdb during huge disk I/O and rebooting system without any sync operation). First constatation : this system has never tried to resynchronized raid. It only tried to run fsck at boot. Second constatation : kernel regulary panics in vfs/ffs subsystem on second workstation (both disks don't contain same data). It never panics on first one. I suspect your explanation is correct. System tries to write on ataraid and write operation aborts as kernel panics. On next reboot, ataraid doesn't resynchronize data. Regards, JKB
Re: texlive xelatex not working [ Was fontconfig : .so.2 not provided by 2.11.1?]
What is preventing you from using the tex packages in pkgsrc? AFAIK, they are based on texlive. On Thu, Jan 21, 2016 at 4:05 AM Mayureshwrote: > On Wed, Jan 20, 2016 at 08:43:33PM +0100, Martin Husemann wrote: > > > amd64 xetex (from texlive) is now looking for freetype.7 and not > finding > > > it. > > > > Then xetex has been compiled against the wrong libs. > > Possibly when the distribution was made, NetBSD 6.x was the latest > version, which might have had freetype.7. > > I'll try recompiling. (I think pkgsrc doesn't host texlive. Will have to > compile outside pkgsrc. Hope, it will be easy as its dependencies are only > on the base.) > > Mayuresh. >
Re: texlive xelatex not working [ Was fontconfig : .so.2 not provided by 2.11.1?]
Mayureshwrites: > On Wed, Jan 20, 2016 at 08:43:33PM +0100, Martin Husemann wrote: >> > amd64 xetex (from texlive) is now looking for freetype.7 and not finding >> > it. >> >> Then xetex has been compiled against the wrong libs. > > Possibly when the distribution was made, NetBSD 6.x was the latest > version, which might have had freetype.7. That should have been clear from the README in the binary distribution (6 vs 7). If not, plase file a bug with them. Note that if you install NetBSD 6, and upgrade to 7, binaries that were compiled against 6 will work fine. Or you can put the old shlibs in place from a 6 distribution. However, you cannot in general successfully build anything on 7 against things that were built for 6, because you will get mixed libraries, > I'll try recompiling. (I think pkgsrc doesn't host texlive. Will have to > compile outside pkgsrc. Hope, it will be easy as its dependencies are only > on the base.) $ ls -ld /usr/pkgsrc/*/*texlive* -rw-r--r-- 1 gdt wheel 1137115288 Jul 22 2010 /usr/pkgsrc/distfiles.wip/texlive-20100722-texmf.tar.xz -rw-r--r-- 1 gdt wheel 179963948 May 30 2013 /usr/pkgsrc/distfiles.wip/texlive-20130530-source.tar.xz drwxr-xr-x 2 gdt users 512 Mar 20 2014 /usr/pkgsrc/distfiles/tex-texlive-docindex-32131 drwxr-xr-x 2 gdt users 512 Jun 2 2015 /usr/pkgsrc/distfiles/tex-texlive-docindex-36749 drwxr-xr-x 2 root users 512 May 5 2010 /usr/pkgsrc/distfiles/tex-texlive-scripts-2009.16483 drwxr-xr-x 2 root users 512 Nov 29 2010 /usr/pkgsrc/distfiles/tex-texlive-scripts-20268 drwxr-xr-x 2 gdt users 512 Nov 27 2013 /usr/pkgsrc/distfiles/tex-texlive-scripts-32115 drwxr-xr-x 2 gdt users 512 Jun 2 2015 /usr/pkgsrc/distfiles/tex-texlive-scripts-36716 drwxr-xr-x 2 gdt users 512 Jun 14 2015 /usr/pkgsrc/distfiles/tex-texlive-scripts-37482 drwxr-xr-x 2 gdt users 512 Nov 27 2013 /usr/pkgsrc/distfiles/tex-texlive.infra-31673 drwxr-xr-x 2 gdt users 512 Mar 9 2015 /usr/pkgsrc/distfiles/tex-texlive.infra-35898 drwxr-xr-x 2 gdt users 512 Jun 14 2015 /usr/pkgsrc/distfiles/tex-texlive.infra-37488 drwxr-xr-x 2 gdt users 512 Sep 1 11:35 /usr/pkgsrc/distfiles/tex-texlive.infra-37738 -rw-r--r-- 1 gdt users 179963948 Nov 13 2013 /usr/pkgsrc/distfiles/texlive-20130530-source.tar.xz -rw-r--r-- 1 gdt users41657688 May 24 2014 /usr/pkgsrc/distfiles/texlive-20140525-source.tar.xz -rw-r--r-- 1 gdt users45459552 May 20 2015 /usr/pkgsrc/distfiles/texlive-20150521-source.tar.xz drwxr-xr-x 3 gdt wheel 512 Nov 30 02:30 /usr/pkgsrc/meta-pkgs/texlive-collection-basic drwxr-xr-x 3 gdt wheel 512 Jun 2 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-basic-doc drwxr-xr-x 3 gdt wheel 512 Sep 1 10:07 /usr/pkgsrc/meta-pkgs/texlive-collection-bibtexextra drwxr-xr-x 3 gdt wheel 512 Sep 1 10:07 /usr/pkgsrc/meta-pkgs/texlive-collection-binextra drwxr-xr-x 3 gdt wheel 512 Jun 14 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-context drwxr-xr-x 3 gdt wheel 512 Sep 3 19:49 /usr/pkgsrc/meta-pkgs/texlive-collection-fontsextra drwxr-xr-x 3 gdt wheel 512 Jun 14 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-fontsrecommended drwxr-xr-x 3 gdt wheel 512 Jun 2 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-fontsrecommended-doc drwxr-xr-x 3 gdt wheel 512 Jun 14 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-fontutils drwxr-xr-x 3 gdt wheel 512 Jun 14 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-formatsextra drwxr-xr-x 3 gdt wheel 512 Jun 2 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-games drwxr-xr-x 3 gdt wheel 512 Dec 14 13:18 /usr/pkgsrc/meta-pkgs/texlive-collection-genericextra drwxr-xr-x 3 gdt wheel 512 Jun 14 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-genericrecommended drwxr-xr-x 3 gdt wheel 512 Jun 2 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-genericrecommended-doc drwxr-xr-x 3 gdt wheel 512 Jun 14 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-htmlxml drwxr-xr-x 3 gdt wheel 512 Jun 14 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-langafrican drwxr-xr-x 3 gdt wheel 512 Sep 1 10:07 /usr/pkgsrc/meta-pkgs/texlive-collection-langarabic drwxr-xr-x 3 gdt wheel 512 Jun 2 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-langchinese drwxr-xr-x 3 gdt wheel 512 Sep 3 19:49 /usr/pkgsrc/meta-pkgs/texlive-collection-langcjk drwxr-xr-x 3 gdt wheel 512 Jun 2 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-langcyrillic drwxr-xr-x 3 gdt wheel 512 Jun 14 2015 /usr/pkgsrc/meta-pkgs/texlive-collection-langczechslovak drwxr-xr-x 3 gdt wheel 512 Sep 3 19:49 /usr/pkgsrc/meta-pkgs/texlive-collection-langenglish drwxr-xr-x 3 gdt wheel 512 Jun 2
Re: wpa_supplicant issues
On January 21, 2016 11:03:38 AM EST, Riccardo Mottolawrote: >hi, that helped a bit. I found that the card was seeing many >accesspoints, but not the one I was trying to connect to. > >I think the culprit is that the access point is "g" only (asked the >system administrator) while my card is running in "b" mode only. > >This is my dmesg: >ipw0 at pci1 dev 2 function 0: vendor 0x8086 product 0x1043 (rev. 0x04) >ipw0: interrupting at irq 11 >ipw0: 802.11 address 00:0c:f1:2e:4a:3b >ipw0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps > >The ThinkPad R51 I own (but got second-hand) came with two adapters as >option: intel 2100 and 2200, only the later is b+g capable. > >Can I know from dmesg or with some other tool from NetBSD which version The information you need is the vendor id and product id, 0x8086 and 0x1043 in this case. A web search for those indicates that you have the 2100 adapter. Eric
Re: ataraid issue was Re: [netbsd-7] Critical issue with ffs+log
Michael van Elst a écrit : joel.bertr...@systella.fr (=?UTF-8?Q?BERTRAND_Jo=c3=abl?=) writes: I suspect your explanation is correct. System tries to write on ataraid and write operation aborts as kernel panics. On next reboot, ataraid doesn't resynchronize data. If both components are online an ataraid mirror will write to both components and read from a component specific to the block number. Then it doesn't really matter if things are synchronized because the component that hasn't been updated isn't used. But ataraid also doesn't support cache flushing. And this can create inconsistent states after a crash. Can you disable write caching on both components ? By dkctl wd0/1 setcache none save ? I will try next saturday. But if both disks are unsynchronized, how can I force a resynchronization ? Regards, JKB
Interesting new hardware that works in 7.0
[Tandberg RDX Removable drive - USB 3.0 model 8666-RDX] http://www.tandbergdata.com/us/index.cfm/products/removable-disk/rdx-quikstor/ I got a this drive for backups. It's basically a rig where they encapsulate SATA 2.5" drives in a hardened case and treat them in a hybrid fashion between tape and disk. The drive itself shows up as a standard USB mass storage device (I got the USB version; there is a SATA version). I get about 14 MB/s write speeds with 'dbench -t 30 16' which is a bit low, but the read speeds are usually between 90-100 MB/s. It really depends on the 'cartridge' (think drop-hardened case around a SATA drive) that you use. If you use an SSD one, you obviously get faster results. They go up to 2TB so far. I did a cheer when "eject sd0" resulted in a motorized eject of the 'cartridge' (ie.. the SCSI enclosure services thingy is working). UFS + logging on drive worked great. Radeon 290X http://www.amazon.com/Sapphire-Version-PCI-Express-Graphics-11226-00-40G/dp/B00HJOKARI It works! It's fast! I can reboot off a different drive (or dual boot) and play games with it, then reboot and run X11 in gloriously fast native mode. I get a gtkperf score of 1.6 seconds !!! ie.. the 2D is wicked fast. The framebuffer console loads automatically even when booting from a 7.0 CDROM. I've waited for years to have a NetBSD rig that I could dual purpose like this... way to go 7.0... -Swift
Re: How to install/build Go applications?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 21.01.2016 21:01, Kamil Rytarowski wrote: > How to build and install Go applications using Go from pkgsrc? > > I mean, not to add new packages, but for example is it possible to > build gogs [1]? > > Thanks > > [1] Gogs (Go Git Service) is a painless self-hosted Git service. > https://gogs.io > For the reference, after fixing ssl bug it works now. http://mail-index.netbsd.org/pkgsrc-bugs/2016/01/21/msg058681.html "pkg/50690: lang/go invalid path to ca-certificates.crt" -BEGIN PGP SIGNATURE- Version: GnuPG v2 iQIcBAEBCAAGBQJWoVI0AAoJEEuzCOmwLnZsJSoP/3h0LBUa47QGA8vfSHZ53u1S r4BS5TodhZEDoyc5JyGwRC6IiADi8Q3oBTGf/QXXfw8zbMit5b8dcwZg0mXCHV+C zyYK969YeO2/1hpCO3Ej8xfE/OHWxss3IUBWk3SrgQwmhfX2hq485WEbbp7Oc7jn zLHCCRwj3cHlmkZTwHvN04SKcRBULiH2dqFjdO7qR4oQrDunyOzAdzCQZu/dfXIN g/ZirNgBkz/mUp81I3WqHeEt7mFfW6A8nREmjp23CPQuxb87WT+d1rZc+k6Ahc6I odzoZbIdH00mKnihzYyGFscPSZrWN+KBVonbgmEtnuwqbm7nIC8swXSBlMJxJuRG 9MP0giz7kVDrZXxxkwCN1fI5cUIz/z9EQ4S2g8tvfl42L5tTq0TJe0kEzvP10T57 ZGE96X9I5147iAOscyflvLMzxoXBb6NOa8VA8ZA/Ide7hDAiqfGf/AVVeualPnKS UaWx1tbQfKEPpRDlfKFl0FnzaWaJPKXQ9/MOJH3uNhi3GB9Pf9sDkxUNHfkn3HEs GKXwHtPpIlaCC9viDMJI6kAhRv3Pbq6gu0H6AhhWXCgUxy12i2DsteSDnndWNKHn PGjmhdv7rfMm9W7Wq/dYHbwxR+Z5odGHSZw95X7xvuCADmehd+pL7Ub5UA4WCJu0 Ui+igaC5SsniltZMlfM6 =0Qd8 -END PGP SIGNATURE-
tmux overhead
Hi, all, Since tmux is part of a standard NetBSD install, I've been using it instead of screen. So far, so good. However, I've noticed that it accumulates a LOT of CPU time even when the underlying tty and shell are doing absolutely nothing. On a 60 MHz m68060 system that's been up for 53 days, for instance, and with a shell that's sitting at a quiet prompt: 25403 john 850 6144K 2312K kqueue44.1H 4.05% 4.05% tmux 44 HOURS? Wow. And while it's otherwise idle, it's taking around 4% CPU. Ideas about why this is so busy? John
Re: tmux overhead
On Thu, 21 Jan 2016, John Klos wrote: Ideas about why this is so busy? No. However, you could try to profile it to find out. Run it in ktrace(1) and/or do a gdb -attach to it's PID. Then start looking at the trace output to see what kind of library functions it's running in it's main event loop. Use kdump to view the ktrace output. That may give you some idea. If you've already tried that... there is always fprintf() logging statements stuffed in the code in key places to see what bits of logic are getting the most action. Swift
Any Postfix solution for mail server migration?
How can this be done? I've Googled and there only seem to be vague hints, but no solid method to migrate in the way I'm proposing. ---[ Problem Description ]--- Server "Oldserv" = old MTA system I want to get away from Server "Newserv" = new MTA I want to migrate TO. 1. Mail to flow to Newserv first. 2. If mail is for a local user, the mail stays on the system. 3. If the mail is for an unknown user it's forwarded to Oldserv. - I don't want the header getting re-written. So, I see people doing things like setting up aliases so that userp...@mydomain.com gets changed to userp...@oldserv.mydomain.com. That's unacceptable. It'll hork up the user's mail experience and re-write addresses that were fine the way they are. The whole point of this is creating a method where I can migrate one user at a time to the new system without having to do a "big bang" migration of the whole company at once. The logistics of that are too big for me. So, I have to be able to take it on in chunks. Plus, I'm not married to Postfix. I'd use Exim or anything else if it'd work. The postfix mailing list and IRC channel are super-hostile & abusive. I certainly wouldn't be heartbroken to never have to communicate with them. This is one of many reasons why I run NetBSD rather than OpenBSD ... Community attitude/spirit also matters to me. This is a simple "checkbox" feature in Communigate Pro, but I'd rather stick with open source, given a choice. Swift
How to install/build Go applications?
How to build and install Go applications using Go from pkgsrc? I mean, not to add new packages, but for example is it possible to build gogs [1]? Thanks [1] Gogs (Go Git Service) is a painless self-hosted Git service. https://gogs.io
Re: Any Postfix solution for mail server migration?
On Thu, Jan 21, 2016 at 02:36:25PM -0700, Swift Griggs wrote: > > How can this be done? I've Googled and there only seem to be vague hints, > but no solid method to migrate in the way I'm proposing. > > ---[ Problem Description ]--- > Server "Oldserv" = old MTA system I want to get away from > Server "Newserv" = new MTA I want to migrate TO. > > 1. Mail to flow to Newserv first. > 2. If mail is for a local user, the mail stays on the system. > 3. If the mail is for an unknown user it's forwarded to Oldserv. > - > > I don't want the header getting re-written. So, I see people doing things > like setting up aliases so that userp...@mydomain.com gets changed to > userp...@oldserv.mydomain.com. That's unacceptable. It'll hork up the user's > mail experience and re-write addresses that were fine the way they are. If it's in the alias file this doens't rewrite the headers, only the envelope. The mail in the final mailbox will still have the original To: or Cc: -- Manuel BouyerNetBSD: 26 ans d'experience feront toujours la difference --
Re: tmux overhead
On 1/21/16 4:59 PM, John Klos wrote: > Ideas about why this is so busy? Hi, John. It might be worth looking into what version of tmux is in the NetBSD version you have installed; maybe it has inefficiencies? Ted Unangst has a blog post at [1] titled "rough idling" where he talks about the situation you're seeing in general and also mentions that a number of efficiency fixes were made to the tmux in OpenBSD after the OpenBSD 5.8 release. I don't know if the same fixes made it into the version of tmux you have; maybe it's worth a look. Regards, Lewis [1] http://www.tedunangst.com/flak/post/rough-idling
Re: Any Postfix solution for mail server migration?
> 3. If the mail is for an unknown user it's forwarded to Oldserv. If "forwarded" means store and forward, you don't want to do that. Consider what happens with bogus names, say from a typo. Your ew system is stuck with bounce or drop. If you aren't familiar with that mess, google for bounce vs reject. I think the solution is to make a list of all the legitimate users on the remote system and setup a forwarding table. It shouldn't be hard, just a bit of scripting. -- These are my opinions. I hate spam.
Re: tmux overhead
jlm...@imca-cat.org ("J. Lewis Muir") writes: >version you have installed; maybe it has inefficiencies? Ted Unangst >has a blog post at [1] titled "rough idling" where he talks about the >situation you're seeing in general and also mentions that a number of >efficiency fixes were made to the tmux in OpenBSD after the OpenBSD 5.8 >release. The window name is refreshed twice a second: 19193 1 tmux 1453438853.358587124 __kevent50(0x3, 0x7bc03000, 0, 0x7bc03800, 0x40, 0x7fffc330) = 0 update_time_cache() (in libevent) { 19193 1 tmux 1453438853.358604936 __clock_gettime50(0x3, 0x7fffc330) = 0 19193 1 tmux 1453438853.358613894 __gettimeofday50(0x7fffc340, 0) = 0 } 19193 1 tmux 1453438853.358643738 __gettimeofday50(0x7fffc2f8, 0) = 0 19193 1 tmux 1453438853.358655873 __gettimeofday50(0x7fffc300, 0) = 0 gethostname() { 19193 1 tmux 1453438853.358713008 __sysctl(0x7fffc208, 0x2, 0x7fffc228, 0x7fffc204, 0, 0) = 0 } osdep_get_name() { 19193 1 tmux 1453438853.359059412 __stat50("/dev/pts/6", 0x7fffc228) = 0 19193 1 tmux 1453438853.359079933 ioctl(0xb, TIOCGPGRP, 0x7fffc1e4) = 0 "n<\0\0" 19193 1 tmux 1453438853.359247015 __sysctl(0x7fffc210, 0x6, 0, 0x7fffc20c, 0, 0) = 0 19193 1 tmux 1453438853.359288681 mmap(0, 0x40, 0x3, 0x14001002, 0x, 0, 0, 0) = 0x7b80 19193 1 tmux 1453438853.359530555 __sysctl(0x7fffc210, 0x6, 0x7b80, 0x7fffc20c, 0, 0) = 0 19193 1 tmux 1453438853.359604513 munmap(0x7b80, 0x40) = 0 } 19193 1 tmux 1453438853.359838573 __gettimeofday50(0x7bc540e0, 0) = 0 19193 1 tmux 1453438853.359870032 __clock_gettime50(0x3, 0x7fffc330) = 0 19193 1 tmux 1453438853.359877896 __gettimeofday50(0x7fffc340, 0) = 0 19193 1 tmux 1453438853.359885032 __clock_gettime50(0x3, 0x7fffc330) = 0 19193 1 tmux 1453438853.359892740 __gettimeofday50(0x7fffc340, 0) = 0 netbsd-current got a new version about two weeks ago that should behave better. -- -- Michael van Elst Internet: mlel...@serpens.de "A potential Snark may lurk in every tree."