Re: Splitting xbaseXY.tgz - stupid idea?
On Sat, May 20, 2006 at 02:43:36AM +0200, viq wrote: Sorry if it sounds otherwise, I have no intention of telling anyone what to do and how, just sharing some idea I had that could possibly satisfy both sides of the argument, and maybe allow to avoid bi-weekly reocurring question. Seeing all those why can't I compile port XX? install xbase but I don't want to install X on my firewall/server/whatever arguments - maybe it would be possible to split xbase into xbase and xlibs packages, with the latter having just some base libraries? I wonder, if xbase were a port, would there have ever been a complaint? what I mean is, if 'make package' or pkg_add just worked, would anyone who has complained have even noticed/cared that xbase got installed? it seems that at least a few people who have complained are perfectly happy installing other stuff they don't really need. no, I'm not suggesting that xbase be a port; I'm just offering some perspective. as far as biweekly question, that should be a clue that the people asking the question aren't doing their homework/paying attention (i.e. they probably would not have noticed/cared if xbase had been installed automatically anyway.) as far as making a new install set, that's a lot of continual work for very little gain. not to mention, it and would add more bytes of text to the installation scripts :( -- [EMAIL PROTECTED]
Re: Splitting xbaseXY.tgz - stupid idea?
On 19 May 2006, Jacob Meuser [EMAIL PROTECTED] wrote: On Sat, May 20, 2006 at 02:43:36AM +0200, viq wrote: Sorry if it sounds otherwise, I have no intention of telling anyone what to do and how, just sharing some idea I had that could possibly satisfy both sides of the argument, and maybe allow to avoid bi-weekly reocurring question. Seeing all those why can't I compile port XX? install xbase but I don't want to install X on my firewall/server/whatever arguments - maybe it would be possible to split xbase into xbase and xlibs packages, with the latter having just some base libraries? I wonder, if xbase were a port, would there have ever been a complaint? what I mean is, if 'make package' or pkg_add just worked, would anyone who has complained have even noticed/cared that xbase got installed? it seems that at least a few people who have complained are perfectly happy installing other stuff they don't really need. I have a simpler question: is there any plan to make installing xbase a requirement in the foreseeable future? no, I'm not suggesting that xbase be a port; I'm just offering some perspective. as far as biweekly question, that should be a clue that the people asking the question aren't doing their homework/paying attention (i.e. they probably would not have noticed/cared if xbase had been installed automatically anyway.) as far as making a new install set, that's a lot of continual work for very little gain. not to mention, it and would add more bytes of text to the installation scripts :( So what you're saying here is that installing 30MB of xbase without the user requesting it is acceptable, but making an install script some 30 bytes larger isn't, right? Regards, Liviu Daia -- Dr. Liviu Daia http://www.imar.ro/~daia
Re: Splitting xbaseXY.tgz - stupid idea?
On Sat, May 20, 2006 at 10:09:15AM +0300, Liviu Daia wrote: On 19 May 2006, Jacob Meuser [EMAIL PROTECTED] wrote: no, I'm not suggesting that xbase be a port; I'm just offering some perspective. as far as biweekly question, that should be a clue that the people asking the question aren't doing their homework/paying attention (i.e. they probably would not have noticed/cared if xbase had been installed automatically anyway.) as far as making a new install set, that's a lot of continual work for very little gain. not to mention, it and would add more bytes of text to the installation scripts :( So what you're saying here is that installing 30MB of xbase without the user requesting it is acceptable, but making an install script some 30 bytes larger isn't, right? Regards, Liviu Daia Under most circumstances, yes. People are far more inconvenienced when the install floppies cease to work than when something installs 30 MB of, admittedly, not-too-useful binaries. And if you really want to be small, you'd better trim the base system too - Apache and Bind are not necessarily useful, after all. I'm now setting up a firewall. It's a very old box, and the only available disk is a Quantum Bigfoot 2.1 GB [1]. I still have some 500 MB of disk space that does not even have a filesystem on it - and quite a bit of space on the various filesystems, too. Of course, if you're installing on CF or somesuch, it *might* be useful to care about this. Then again, working at the supermarket for a couple of hours will net you the money to buy a bigger card, quite probably in less time than it takes to adequately strip baseXY.tgz and xbaseXY.tgz. Joachim [1] For those who don't know, this disk is not only old, and small, but also *very* slow. It was made in large quantities for the home customer market, where storage capacity count(s/ed) for more than speed. However, as long as it boots, it's fine for a firewall - it's not as if it's doing *anything* except running the occasional backup and storing a select few logs...
Re: XF4.tar.gz in /usr or /usr/src?
On Fri, May 19, 2006 at 09:06:24PM +0200, Tobias Weisserth wrote: Hi everybody, I hope this is the right place to post this. (...) I read the instructions for the second errata (ftp://ftp.openbsd.org/pub/OpenBSD/patches/3.9/common/002_xorg.patch). It reads: Apply by doing: cd /usr/src/XF4 patch -p0 002_xorg.patch And then rebuild and install X: make build This conflicts with what I did according to http://www.openbsd.org/ anoncvs.html: To extract the source tree from the CD to /usr/src (assuming the CD is mounted on /mnt): # cd /usr/src; tar xzf /mnt/src.tar.gz # cd /usr; tar xzf /mnt/XF4.tar.gz # tar xzf /mnt/ports.tar.gz I unpacked XF4.tar.gz in /usr like the web page suggests, but the patch assumes the XF4 sources are located in /usr/src. So I have no / usr/src/XF4 directory. I assume the patch instructions are correct and the web page is wrong? I just moved the XF4 directory into /usr/src and applied the instructions from the patch. It compiled for some time and just as I'm writing this it aborted with multiple error code 1 messages in the Makefile. I guess I misunderstood something here. Can anybody help me out please? I'm a little confused about this. Thanks. Ultimately, it doesn't matter where you keep X. My tree lives under /usr/src/XF4, with a symlink from /usr/XF4 just to be sure. I'm fairly certain both things work; the canonical way, though, is to put XF4 under /usr. Joachim
Error compiling 3.9 -stable kernel
I have extracted the sources from the CDROM archive. Upgraded to 3.9 -stable with: cd /usr cvs -d [EMAIL PROTECTED]:/cvs checkout -P -rOPENBSD_3_9 src/sys And then I tried to compile the GENERIC kernel with the usual procedure, but the following error appeared: cc -Werror -Wall -Wstrict-prototypes -Wmissing-prototypes -Wno-uninitialized -Wno-format -Wno-main -Wno-sign-compare -mcmodel=kernel -mno-red-zone -fno-strict-aliasing -mno-sse2 -mno-sse -mno-3dnow -mno-mmx -msoft-float -fno-builtin-printf -fno-builtin-log -fno-omit-frame-pointer -O2 -pipe -nostdinc -I. -I/usr/src/sys/arch/amd64/compile/GENERIC/../../../../arch -I/usr/src/sys/arch/amd64/compile/GENERIC/../../../.. -DDDB -DDIAGNOSTIC -DKTRACE -DACCOUNTING -DKMEMSTATS -DPTRACE -DCRYPTO -DSYSVMSG -DSYSVSEM -DSYSVSHM -DUVM_SWAP_ENCRYPT -DCOMPAT_35 -DCOMPAT_43 -DLKM -DFFS -DFFS_SOFTUPDATES -DUFS_DIRHASH -DQUOTA -DEXT2FS -DMFS -DXFS -DTCP_SACK -DTCP_ECN -DTCP_SIGNATURE -DNFSCLIENT -DNFSSERVER -DCD9660 -DUDF -DMSDOSFS -DFIFO -DPORTAL -DINET -DALTQ -DINET6 -DIPSEC -DPPP_BSDCOMP -DPPP_DEFLATE -DMROUTING -DBOOT_CONFIG -DUSER_PCICONF -DAPERTURE -DPCIVERBOSE -DUSBVERBOSE -DWSDISPLAY_COMPAT_USL -DWSDISPLAY_COMPAT_RAWKBD -DWSDISPLAY_DEFAULTSCREENS=6 -DWSDISPLAY_COMPAT_PCVT -D_KERNEL -Damd64 -Dx86_64 -c /usr/src/sys/arch/amd64/compile/GENERIC/../../../../dev/i2c/adm1026.c mkdir -p /usr/src/sys/arch/amd64/compile/GENERIC/lib/kern making sure the kern library is up to date... make: don't know how to make ../../machine/types.h. Stop in /usr/src/sys/lib/libkern. *** Error code 2 Stop in /usr/src/sys/arch/amd64/compile/GENERIC (line 28 of /usr/src/sys/arch/amd64/compile/GENERIC/../../../../lib/libkern/Makefile.inc). What could be the problem? Thanks. -- ___ __ |- [EMAIL PROTECTED] |ederico Giannici http://www.neomedia.it ___
Re: Error compiling 3.9 -stable kernel
On Sat, May 20, 2006 at 12:36:16PM +0200, Federico Giannici wrote: make: don't know how to make ../../machine/types.h. Stop in Google for it. And read http://www.openbsd.org/faq/current.html. Ciao, Kili -- In celebration and out of respect for Puffy, we will not be serving sushi in the cafeteria today. -- Bill, 18th oct. 2005 (OpenBSD's 10th birthday)
an easy way to black list IP's
Hi all, I'm using spamd and it does a great job. What I'm trying to figure out is how to easily add the IP's of the sending mail server for the few spam that still get through. By easy, I mean for clients of mine who use Exchange/Outlook, where I put a obsd box running spamd in front of Exchange. I am trying to find a way where I could tell my clients that when some spam does get through, just forward that spam to a particular email address. Some process will extract the IP of the MTA that sent the spam and blacklist it. I installed and played around with relaydb from ports, but that doesn't work with emails that have been forwarded. Any ideas??
Re: Error compiling 3.9 -stable kernel
I'm replying to myself: I extracted the 3.9 source code in another machine only, so I used the old 3.8 one... Please, forgive me. Thanks. Federico Giannici wrote: I have extracted the sources from the CDROM archive. Upgraded to 3.9 -stable with: cd /usr cvs -d [EMAIL PROTECTED]:/cvs checkout -P -rOPENBSD_3_9 src/sys And then I tried to compile the GENERIC kernel with the usual procedure, but the following error appeared: cc -Werror -Wall -Wstrict-prototypes -Wmissing-prototypes -Wno-uninitialized -Wno-format -Wno-main -Wno-sign-compare -mcmodel=kernel -mno-red-zone -fno-strict-aliasing -mno-sse2 -mno-sse -mno-3dnow -mno-mmx -msoft-float -fno-builtin-printf -fno-builtin-log -fno-omit-frame-pointer -O2 -pipe -nostdinc -I. -I/usr/src/sys/arch/amd64/compile/GENERIC/../../../../arch -I/usr/src/sys/arch/amd64/compile/GENERIC/../../../.. -DDDB -DDIAGNOSTIC -DKTRACE -DACCOUNTING -DKMEMSTATS -DPTRACE -DCRYPTO -DSYSVMSG -DSYSVSEM -DSYSVSHM -DUVM_SWAP_ENCRYPT -DCOMPAT_35 -DCOMPAT_43 -DLKM -DFFS -DFFS_SOFTUPDATES -DUFS_DIRHASH -DQUOTA -DEXT2FS -DMFS -DXFS -DTCP_SACK -DTCP_ECN -DTCP_SIGNATURE -DNFSCLIENT -DNFSSERVER -DCD9660 -DUDF -DMSDOSFS -DFIFO -DPORTAL -DINET -DALTQ -DINET6 -DIPSEC -DPPP_BSDCOMP -DPPP_DEFLATE -DMROUTING -DBOOT_CONFIG -DUSER_PCICONF -DAPERTURE -DPCIVERBOSE -DUSBVERBOSE -DWSDISPLAY_COMPAT_USL -DWSDISPLAY_COMPAT_RAWKBD -DWSDISPLAY_DEFAULTSCREENS=6 -DWSDISPLAY_COMPAT_PCVT -D_KERNEL -Damd64 -Dx86_64 -c /usr/src/sys/arch/amd64/compile/GENERIC/../../../../dev/i2c/adm1026.c mkdir -p /usr/src/sys/arch/amd64/compile/GENERIC/lib/kern making sure the kern library is up to date... make: don't know how to make ../../machine/types.h. Stop in /usr/src/sys/lib/libkern. *** Error code 2 Stop in /usr/src/sys/arch/amd64/compile/GENERIC (line 28 of /usr/src/sys/arch/amd64/compile/GENERIC/../../../../lib/libkern/Makefile.inc). What could be the problem? Thanks.
Re: an easy way to black list IP's
On 5/20/06, Craig Hammond [EMAIL PROTECTED] wrote: Hi all, I'm using spamd and it does a great job. What I'm trying to figure out is how to easily add the IP's of the sending mail server for the few spam that still get through. By easy, I mean for clients of mine who use Exchange/Outlook, where I put a obsd box running spamd in front of Exchange. I am trying to find a way where I could tell my clients that when some spam does get through, just forward that spam to a particular email address. Some process will extract the IP of the MTA that sent the spam and blacklist it. I installed and played around with relaydb from ports, but that doesn't work with emails that have been forwarded. Any ideas?? You do know that headers can be forged right? So an automagic forward - |/script - blacklist from a pissed off user can end up blacklisting a legitimate MTA. You may want to just look into greylisting and using some aggressive milters (milter_regex is my savior). Other than that, just read aliases(5), forward(5) or look into procmail
Linux UFS write support ??
Hello, I'm trying to mount a OpenBSD image locally with write support on linux. I've recompiled my kernel to enable this feature. I mount the image by: mount -t ufs -o ufstype=44bsd,loop,rw my_image.fs /mnt It mount well, But if I try: dd if=myfile of=/mnt/myfile bs=512 it freeze and I've to reboot the machine. Nothing in log. If I try a cp ~/.profile /mnt it works, but if do the same cp again, it freeze ... I'm using ubuntu-server 5.10/ Kernel version is 2.6.12 I've tried to do the same manipulation with a real 44bsd partition, and the problem is the same, It's not loop fault. Can anyone help me on this ? Thanks you very much. ++ Jerome [demime 1.01d removed an attachment of type application/x-pkcs7-signature which had a name of smime.p7s]
Re: New server
On 20 May 2006, at 00:44, Stuart Henderson wrote: move the files under /var/www, and nfs mount to 127.0.0.1 back into the homes? you probably want to look at amd for this. of course the ftpd could sit on another machine if you want. This means that I'd need an nfs mount point for each website running on that machine (a lot more than 80), and also requiring the use of nfs. moving the whole homes under /var/www is simpler and presumably more robust, of course... and hey, it's only 80. Which defeats the object of what I'm trying to achieve; user's websites (and only their websites) are inside the apache chroot, so in the event of a php or apache exploit, only their websites are exposed, not their entire home directory or Maildir. Something's got to give here. I suspect that I'm going to have to un- chroot the ftp daemon. Is there an ftpd somewhere that can prevent users from looking at certain directories? For example, I would like to limit access only to /home/username and /var/www/home/username in ftpd, and prevent access to places like /etc, /usr/local, and so on. Gaby -- Junkets for bunterish lickspittles since 1998! http://www.playr.co.uk/sudoku/ http://weblog.vanhegan.net/
Re: Linux UFS write support ??
On Sat, 20 May 2006 15:02:31 +0200, Jirtme Loyet wrote: Hello, I'm trying to mount a OpenBSD image locally with write support on linux. I've recompiled my kernel to enable this feature. I mount the image by: mount -t ufs -o ufstype=44bsd,loop,rw my_image.fs /mnt It mount well, But if I try: dd if=myfile of=/mnt/myfile bs=512 it freeze and I've to reboot the machine. Nothing in log. If I try a cp ~/.profile /mnt it works, but if do the same cp again, it freeze ... I'm using ubuntu-server 5.10/ Kernel version is 2.6.12 I've tried to do the same manipulation with a real 44bsd partition, and the problem is the same, It's not loop fault. Can anyone help me on this ? You are asking in the wrong place. This is not an OpenBSD problem. Why would we know why linux screws up on this task? Perhaps you could ask you distro provider. Thanks you very much. No prob. From the land down under: Australia. Do we look umop apisdn from up over? Do NOT CC me - I am subscribed to the list. Replies to the sender address will fail except from the list-server.
Re: Linux UFS write support ??
You are asking in the wrong place. This is not an OpenBSD problem. Why would we know why linux screws up on this task? Perhaps you could ask you distro provider. I was asking there in case that someone has ever hade this problem. Thx ++ JErome [demime 1.01d removed an attachment of type application/x-pkcs7-signature which had a name of smime.p7s]
Re: Linux UFS write support ??
On 20 May 2006, Jirtme Loyet [EMAIL PROTECTED] wrote: Hello, I'm trying to mount a OpenBSD image locally with write support on linux. [...] Don't do this; it will trash your filesystem. While read-only UFS on Linux is relatively safe these days (it used to produce frequent kernel panics), read-write has never worked properly. I also doubt there is much interest in fixing it. Regards, Liviu Daia -- Dr. Liviu Daia http://www.imar.ro/~daia
nmap 3.95/4.03 core dumps on OpenBSD 3.9 if -T[0,1] was used
That was mailed to nmap-dev and the Portmaintainer with no reply so far: Well I asked already for an Update for OpenBSD 3.9 STABLE but nmap 4.03 is just avaiable for current... Anyway there`s another issue (wich is NOT related to mem-leaks in 3.95 wich make nmap core-dump anyway): Problem: sudo nmap -P0 -T1 -sV -vvv -oA output 10.10.128-143.* 3.95: Initiating ARP Ping Scan against 1225 hosts [1 port/host] at 12:21 ARP Ping Scan Timing: About 0.41% done; ETC: 14:25 (2:03:13 remaining) assertion 0 failed: file scan_engine.cc, line 1826, function ultrascan_port_pspec_update Abort trap (core dumped) So... (for OpenBSD) export [EMAIL PROTECTED]:/cvs cd /tmp cvs get ports/net/nmap cd ports/net/nmap sudo pkg_delete nmap* sudo env FLAVOR=no_x11 make install Starting Nmap 4.03 ( http://www.insecure.org/nmap/ ) at 2006-05-15 12:28 CEST Initiating ARP Ping Scan against 1225 hosts [1 port/host] at 12:28 ARP Ping Scan Timing: About 0.41% done; ETC: 14:32 (2:03:13 remaining) assertion 0 failed: file scan_engine.cc, line 1683, function ultrascan_port_pspec_update Abort trap (core dumped) Are there any problems... with the Timing-Settings? If I use f.e. -T[2,3,4,5] (or I simply do not -T) it works... Starting Nmap 4.03 ( http://www.insecure.org/nmap/ ) at 2006-05-15 12:34 CEST Initiating ARP Ping Scan against 1225 hosts [1 port/host] at 12:34 The ARP Ping Scan took 1.46s to scan 1225 total hosts. DNS resolution of 1042 IPs took 13.12s. Mode: Async [#: 1, OK: 938, NX: 37, DR: 67, SF: 0, TR: 1361, CN: 0] Initiating SYN Stealth Scan against 5 hosts [1674 ports/host] at 12:34 Discovered open port 22/tcp on 10.10.128.6 Discovered open port 22/tcp on 10.10.128.7 Discovered open port 443/tcp on 10.10.128.6 Discovered open port 443/tcp on 10.10.128.7 Discovered open port 53/tcp on 10.10.128.6 Discovered open port 80/tcp on 10.10.128.6 Discovered open port 25/tcp on 10.10.128.6 Discovered open port 21/tcp on 10.10.128.6 Discovered open port 3306/tcp on 10.10.128.6 Discovered open port 465/tcp on 10.10.128.6 Discovered open port 8443/tcp on 10.10.128.6 Discovered open port 993/tcp on 10.10.128.6 Discovered open port 111/tcp on 10.10.128.7 Discovered open port 143/tcp on 10.10.128.6 Discovered open port 995/tcp on 10.10.128.6 Discovered open port 106/tcp on 10.10.128.6 Discovered open port 110/tcp on 10.10.128.6 Btw: Would it be possible to add --debug to ./configure? Or for the Portmaintainer: a Debug-Flavor maybe? Well I rebuild it with debugging infos anyway: nmap 4.03: gdb -c nmap.core ./nmap Core was generated by `nmap'. Program terminated with signal 6, Aborted. Reading symbols from /usr/local/lib/libpcre.so.1.0...done. Loaded symbols for /usr/local/lib/libpcre.so.1.0 Reading symbols from /usr/lib/libpcap.so.4.0...done. Loaded symbols for /usr/lib/libpcap.so.4.0 Reading symbols from /usr/lib/libssl.so.10.0...done. Loaded symbols for /usr/lib/libssl.so.10.0 Reading symbols from /usr/lib/libcrypto.so.12.0...done. Loaded symbols for /usr/lib/libcrypto.so.12.0 Reading symbols from /usr/local/lib/libdnet.so.1.0...done. Loaded symbols for /usr/local/lib/libdnet.so.1.0 Reading symbols from /usr/lib/libstdc++.so.42.0...done. Loaded symbols for /usr/lib/libstdc++.so.42.0 Reading symbols from /usr/lib/libm.so.2.1...done. Loaded symbols for /usr/lib/libm.so.2.1 Reading symbols from /usr/lib/libc.so.39.0...done. Loaded symbols for /usr/lib/libc.so.39.0 Reading symbols from /usr/libexec/ld.so...done. Loaded symbols for /usr/libexec/ld.so #0 0x0bcf4995 in kill () from /usr/lib/libc.so.39.0 I requested an Update to nmap 4.03 for OpenBSD 3.9 STABLE because 3.95 has a mem-leak wich makes it (sometimes...) crash. I use 4.03 on 39 STABLE because 3.95 is for me a littlebit useless if it crashs once (or more) a day. :-/ Kind regrds, Sebastian
Re: an easy way to black list IP's
On Sat, May 20, 2006 at 09:49:31AM -0400, Jim Razmus wrote: Take a look at mail/relaydb in the ports tree. Also check the archives as this has been discussed at depth and included several solutions. Why ports instead of packages? Notably since he's already tried relaydb (and it doesn't do forwarded messages). Are there other flavours of interest? -- Darrin Chandler| Phoenix BSD Users Group [EMAIL PROTECTED] | http://bsd.phoenix.az.us/ http://www.stilyagin.com/ |
Re: New server
On Sat, May 20, 2006 at 02:14:34PM +0100, Gaby vanhegan wrote: On 20 May 2006, at 00:44, Stuart Henderson wrote: move the files under /var/www, and nfs mount to 127.0.0.1 back into the homes? you probably want to look at amd for this. of course the ftpd could sit on another machine if you want. This means that I'd need an nfs mount point for each website running on that machine (a lot more than 80), and also requiring the use of nfs. moving the whole homes under /var/www is simpler and presumably more robust, of course... and hey, it's only 80. Which defeats the object of what I'm trying to achieve; user's websites (and only their websites) are inside the apache chroot, so in the event of a php or apache exploit, only their websites are exposed, not their entire home directory or Maildir. Something's got to give here. I suspect that I'm going to have to un- chroot the ftp daemon. Is there an ftpd somewhere that can prevent users from looking at certain directories? For example, I would like to limit access only to /home/username and /var/www/home/username in ftpd, and prevent access to places like /etc, /usr/local, and so on. A lot of FTP daemons can do that, but I don't really see the point. The protections they offer might or might not be circumventable, but nothing interesting should be readable anyway. Anyway, ISTR that ProFTPd could do that; I'm quite certain neither stock ftpd nor vsftpd can. Joachim
LSI MegaRaid non-hotspare
Hi, As mentioned before, I have a new server with the LSI MegaRaid SATA150-4 card. All works nicely at the moment, bar a slight problem with hot-spares. We configured a RAID-5 array with three 250Gb drives and one hot spare. We simulated a failure by yanking the cable out from drive 2, and the alarm went off, bioctl allowed us to silence it, and showed that the array was rebuilding, onto disk 3. The rebuild process took about 9 hours (64bit card in a 32bit slot). We put the drive back in, and bioctl showed the drive as Unused. So we try to promote that drive back to a hot spare, but the bioctl command: # bioctl -H 0:2.0 ami0 Seems to return nothing, nor does it make the change. We tried rebooting, but there's no change, and the command still does the same. When we boot into the MegaRaid config utility on the card's BIOS, it shows the drive as a hot spare, whereas bioctl still reports it as unused. # bioctl -Dhiv ami0 bioctl: cookie = 0xd2882ca0 bio_inq Volume Status Size Device ami0 0 Online 468G sd0 RAID5 0 Online 234G 0:0.0 noencl Maxtor 6V250F0 VA11 'V5075JFG' 1 Online 234G 0:1.0 noencl Maxtor 6V250F0 VA11 'V5075JVG' 2 Online 234G 0:3.0 noencl Maxtor 6V250F0 VA11 'V5064EEG' ami0 1 Unused 234G 0:2.0 noencl Maxtor 6V250F0 VA11 'V5075LQG' # bioctl -Dhiv -H 0:2.0 ami0 bioctl: cookie = 0xd2882ca0 bio_inq Volume Status Size Device ami0 0 Online 468G sd0 RAID5 0 Online 234G 0:0.0 noencl Maxtor 6V250F0 VA11 'V5075JFG' 1 Online 234G 0:1.0 noencl Maxtor 6V250F0 VA11 'V5075JVG' 2 Online 234G 0:3.0 noencl Maxtor 6V250F0 VA11 'V5064EEG' ami0 1 Unused 234G 0:2.0 noencl Maxtor 6V250F0 VA11 'V5075LQG' Any suggestions? In order to get the kernel to boot we had to disable pcibios using config, which we did on a copy of bsd.mp. We took a backup of the fresh bsd.mp. Here's a dmesg: OpenBSD 3.9 (GENERIC.MP) #598: Thu Mar 2 02:37:06 MST 2006 [EMAIL PROTECTED]:/usr/src/sys/arch/i386/compile/GENERIC.MP cpu0: Intel(R) Pentium(R) D CPU 2.66GHz (GenuineIntel 686-class) 2.68 GHz cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36, CFLUSH,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,TM2,CNXT-ID real mem = 2146541568 (2096232K) avail mem = 1952505856 (1906744K) using 4278 buffers containing 107429888 bytes (104912K) of memory mainbus0 (root) bios0 at mainbus0: AT/286+(00) BIOS, date 10/30/05, BIOS32 rev. 0 @ 0xf0010 apm0 at bios0: Power Management spec V1.2 apm0: AC on, battery charge unknown apm0: flags 30102 dobusy 0 doidle 1 pcibios at bios0 function 0x1a not configured bios0: ROM list: 0xc/0x8000 0xc8000/0x2200 mainbus0: Intel MP Specification (Version 1.1) (INTELPremium ) cpu0 at mainbus0: apid 0 (boot processor) cpu0: apic clock running at 133 MHz cpu1 at mainbus0: apid 1 (application processor) cpu1: Intel(R) Pentium(R) D CPU 2.66GHz (GenuineIntel 686-class) 2.68 GHz cpu1: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36, CFLUSH,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,SSE3,MWAIT,TM2,CNXT-ID mainbus0: bus 0 is type PCI mainbus0: bus 1 is type PCI mainbus0: bus 2 is type PCI mainbus0: bus 3 is type PCI mainbus0: bus 4 is type PCI mainbus0: bus 5 is type ISA ioapic0 at mainbus0: apid 2 pa 0xfec0, version 20, 24 pins pci0 at mainbus0 bus 0: configuration mode 1 (no bios) pchb0 at pci0 dev 0 function 0 Intel 82955X MCH rev 0x81 ppb0 at pci0 dev 28 function 0 Intel 82801GB PCIE rev 0x01 pci1 at ppb0 bus 4 ppb1 at pci0 dev 28 function 4 Intel 82801G PCIE rev 0x01 pci2 at ppb1 bus 3 em0 at pci2 dev 0 function 0 Intel PRO/1000MT (82573L) rev 0x00: apic 2 int 16 (irq 11), address 00:15:f2:c8:8e:10 ppb2 at pci0 dev 28 function 5 Intel 82801G PCIE rev 0x01 pci3 at ppb2 bus 2 CMD Technology SiI3132 SATA rev 0x01 at pci3 dev 0 function 0 not configured uhci0 at pci0 dev 29 function 0 Intel 82801GB USB rev 0x01: apic 2 int 20 (irq 10) usb0 at uhci0: USB revision 1.0 uhub0 at usb0 uhub0: Intel UHCI root hub, rev 1.00/1.00, addr 1 uhub0: 2 ports with 2 removable, self powered uhci1 at pci0 dev 29 function 1 Intel 82801GB USB rev 0x01: apic 2 int 17 (irq 10) usb1 at uhci1: USB revision 1.0 uhub1 at usb1 uhub1: Intel UHCI root hub, rev 1.00/1.00, addr 1 uhub1: 2 ports with 2 removable, self powered
Re: Raid 1 and 2 Disks: kernel panic with init: not found when reboot into broken mirror
On Wed, May 17, 2006 at 12:39:57AM +0200, ip wrote: On 5/14/06, Joachim Schipper [EMAIL PROTECTED] wrote: While wd1a does have a kernel, it does not have a proper root filesystem - for instance, no /dev directory, or more specifically no /dev/console. Fix this, and also have a look at daily(8) which documents the altroot mechanism, which is quite useful to ensure backup kernels can always be found in a RAIDed system. Joachim Hello misc, Hi Joachim and thanks for the tips. However, I don't understand why I receive this errors. From the raidctl man page: Section: Auto-configuration and Root on RAID ... RAID sets which are auto-configurable will be configured before the root file system is mounted. These RAID sets are thus available for use as a root file system, or for any other file system. [snip] Note that kernels can't be directly read from a RAID component. To sup- port the root file system on RAID sets, some mechanism must be used to get a kernel booting. For example, a small partition containing only the secondary boot-blocks and an alternate kernel (or two) could be used. Once a kernel is booting however, and an auto-configured RAID set is found that is eligible to be root, then that RAID set will be auto-con- figured and its `a' partition (aka raid[0..n]a) will be used as the root file system. ... So, when I make the wd1a partition, I think that bsd and boot files are sufficient for the goal. Instead, during the reboot into degrade mode, the error messages seem to indicate that the Auto RAID system has not been activated. Infact I have fixed dev/console and other mechanisms, but this solution go to recreate a minimum complete installation into wd1a... (Sorry for the slow reaction, but the OP might not have this one figured out yet, or at least it'll be useful for the archives...) When you booted, and this was in your original message you snipped above, you posted that you received: ... Kernelized RAIDFrame activated dkcsum wd0 matches BIOS drive 0x80 dkcsum wd1 matches BIOS drive 0x81 root on wd1a rootdev=0x10 rrootdev=0x320 rawdev=0x312 warning: /dev/console does not exist init: not found panic: no init ... ddb This looks like the RAID is not autoconfiguring at all. Maybe you didn't build the kernel with the proper options, or maybe you forgot raidctl -A root? Joachim
Re: New server
On 20 May 2006, at 15:15, Joachim Schipper wrote: Something's got to give here. I suspect that I'm going to have to un- chroot the ftp daemon. Is there an ftpd somewhere that can prevent users from looking at certain directories? For example, I would like to limit access only to /home/username and /var/www/home/username in ftpd, and prevent access to places like /etc, /usr/local, and so on. A lot of FTP daemons can do that, but I don't really see the point. The protections they offer might or might not be circumventable, but nothing interesting should be readable anyway. If the ftpd runs as the UID of the person that's logged in, they won't be able to access the files they don't own anyway (contents of / etc, and others). But if possible, I'd just like to hide them from view, so they can't even be read. For example, # ls -lFa /etc | grep passwd -rw--- 1 root wheel 2688 May 19 21:57 master.passwd -rw-r--r-- 1 root wheel 2235 May 19 21:57 passwd Would still result in somebody with FTP access being able to download a list of users on the system. I would like to prevent them from doing that if possible. Anyway, ISTR that ProFTPd could do that; I'm quite certain neither stock ftpd nor vsftpd can. I hear that the security record of ProFTPd is not stellar, to say the least. I'm fairly sure that the stock ftpd can't, and I can't find anything in pure-ftpd about it either. Gaby -- Junkets for bunterish lickspittles since 1998! http://www.playr.co.uk/sudoku/ http://weblog.vanhegan.net/
ksh: typeset screwing up subsequent parameter's array indices?
=== [/home/jrrs] $ uname -mrpsv OpenBSD 3.9 GENERIC.MP#690 i386 AMD Athlon(tm) MP 2800+ (AuthenticAMD 686-class, 512KB L2 cache) [/home/jrrs] $ echo $KSH_VERSION @(#)PD KSH v5.2.14 99/07/13.2 [/home/jrrs] $ ls -l $(which ksh) -r-xr-xr-x 3 root bin 324128 May 1 20:28 /bin/ksh === if i start a new shell ( ie: new screen(1) window ), it almost looks like if i typeset a variable right, ( wrong? basically use -L1 and 0x000[0-9] ) and set about to make others. the name of the next variable gets tainted for the duration of that shell's existance (even if i unset the one i made with typeset -L1 right after i make it) plz let me just show it below: ( in specific, i'm puzzled by why when i make something an array by doing a VAR[index] assignment, it makes the original parameter in the VAR have an index of 805384193 instead of 0. ). if it has any significance, (2**30+2**29)/2 is quite close to that 805384193 number. i took a few stabs at the number but didn't unravel it neatly into some 'powers of two' constituents yet === $ typeset -L1 __A=0x0009 $ typeset -Ui10 __B=1 $ __B[2]=990 $ integer | grep ^__ __B[805384193]=1 __B[2]=990 $ unset __B $ typeset -Ui10 __B=1 $ __B[2]=990 $ integer | grep ^__ __B[805384193]=1 __B[2]=990 $ unset __A $ unset __B $ typeset -Ui10 __B=1 $ __B[2]=990 $ integer | grep ^__ __B[805384193]=1 __B[2]=990 $ unset __B $ typeset | grep ^__ $ JOE=9 $ JOE[2]=9 $ set | grep JOE JOE[0]=9 JOE[2]=9 $ integer MAILCHECK=600 OPTIND=1 PPID=22746 RANDOM SECONDS=164 TMOUT=3600 $ typeset -i J=50 $ integer J=50 MAILCHECK=600 OPTIND=1 PPID=22746 RANDOM SECONDS=175 TMOUT=3600 $ J[1]=55 $ integer J[0]=50 J[1]=55 MAILCHECK=600 OPTIND=1 PPID=22746 RANDOM SECONDS=183 TMOUT=3600 $ typeset -i __B=50 $ __B[2]=55 $ integer J[0]=50 J[1]=55 MAILCHECK=600 OPTIND=1 PPID=22746 RANDOM SECONDS=235 TMOUT=3600 __B[805384193]=50 __B[2]=55 == even though i unset the __B and it appears it was missing, when made it again, it got all ho'd up when i did the array thing closed that, started a new shell, tried things slightly different, got what seems to be same result === $ typeset -L1 A=0x000a $ A=''; B=''; C=''; M='' $ C[2]=38; B[5]=93; M[9]=aaa; A[3]=ac93u $ set | grep ^.\\\[ A[0]=' ' A[3]=a B[805384193]= B[5]=93 C[0]= C[2]=38 M[0]= M[9]=aaa === duplicated with ssh [EMAIL PROTECTED]: === Terminal type? [screen] vt100 # typeset -L1 P=0x0008 # typeset -i O=900 # O[2]=83 # integer MAILCHECK=600 O[805384193]=900 O[2]=83 OPTIND=1 PPID=11522 RANDOM SECONDS=31 TMOUT=0 === thus far, any time i see something get 805384193'd, that param name is toasted and will always keep getting 805384193'd if i make it an array. unsetting a param that was 8-3'd, assigning it as a string and then testing, same. look at this poor bastard. this was after trying to set the param back to string: === $ POOP[1]=lala $ set | grep POOP POOP[805384193]=a POOP[1]=lala [/home/jrrs] $ unset POOP [/home/jrrs] $ set | grep POOP _=POOP [/home/jrrs] $ POOP[1]=lala [/home/jrrs] $ set | grep POOP POOP[1]=lala _=POOP [/home/jrrs] $ POOP=poop [/home/jrrs] $ set | grep POOP POOP[805384193]=poop POOP[1]=lala _=POOP === *if* i set it explicltly to something[0]=blahblah, it seems that corrects the 0th array index to actually be 0, for that and all further instances, but if i don't zero it out, it stays at 805blahblah3. i have half a mind that says otto@/kili@/ckuthe will come down and cluestick me about why what i am seeing is expected behaviour, but the other half expects this to be not what should be happening. -- jared [ openbsd 3.9-current GENERIC ( may 1 ) // i386 ]
Re: LSI MegaRaid non-hotspare
On 20 May 2006, at 16:28, Marco Peereboom wrote: I fixed this in current. You can simply just upgrade the ami files to -current and build a 3.9 that is mostly RELEASE. Was it a functional problem or just a cosmetic one? If I leave it as it is, is it going to cause any real problems for me? Gaby -- Junkets for bunterish lickspittles since 1998! http://www.playr.co.uk/sudoku/ http://weblog.vanhegan.net/
Promise SATA 300 TX4.
Hi all! Is there anyone out there using this controller successfully with OpenBSD ? In other word's : Is it supported by this OS ? /Hans Almqvist
Re: New server
At 09:14 AM 5/20/2006, you wrote: On 20 May 2006, at 00:44, Stuart Henderson wrote: move the files under /var/www, and nfs mount to 127.0.0.1 back into the homes? you probably want to look at amd for this. of course the ftpd could sit on another machine if you want. This means that I'd need an nfs mount point for each website running on that machine (a lot more than 80), and also requiring the use of nfs. moving the whole homes under /var/www is simpler and presumably more robust, of course... and hey, it's only 80. Which defeats the object of what I'm trying to achieve; user's websites (and only their websites) are inside the apache chroot, so in the event of a php or apache exploit, only their websites are exposed, not their entire home directory or Maildir. Something's got to give here. I suspect that I'm going to have to un- chroot the ftp daemon. Is there an ftpd somewhere that can prevent users from looking at certain directories? For example, I would like to limit access only to /home/username and /var/www/home/username in ftpd, and prevent access to places like /etc, /usr/local, and so on. Gaby I use Pro FTP to chroot users to their home directories. see http://www.proftpd.org/
Re: LSI MegaRaid non-hotspare
As mentioned before, I have a new server with the LSI MegaRaid SATA150-4 card. All works nicely at the moment, bar a slight problem with hot-spares. We configured a RAID-5 array with three 250Gb drives and one hot spare. We simulated a failure by yanking the cable out from drive 2, and the alarm went off, bioctl allowed us to silence it, and showed that the array was rebuilding, onto disk 3. The rebuild process took about 9 hours (64bit card in a 32bit slot). We put the drive back in, and bioctl showed the drive as Unused. So we try to promote that drive back to a hot spare, but the bioctl command: # bioctl -H 0:2.0 ami0 Seems to return nothing, nor does it make the change. We tried rebooting, but there's no change, and the command still does the same. When we boot into the MegaRaid config utility on the card's BIOS, it shows the drive as a hot spare, whereas bioctl still reports it as unused. Right. The card honours your request for the device to be a hot spare, but something was busted in reporting the new hot spare. Apparently this bug is now fixed: revision 1.156 date: 2006/05/12 20:51:25; author: marco; state: Exp; lines: +14 -22 Fix a misreporting bug after bioctl -H is used to create a hotspare. This was reported by several people. What happens is that the firmware sometimes misreports what SCSI type a device is. The driver was only allowing a create hotspare function when the type was set to hard disk. Since the firmware will, obviously, not allow the driver to create a hotspare on any other type of device the driver doesn't need these smarts and now will ignore the type. Tested by henning todd and Ben Lovett. ok dlg
Re: New server
On 20 May 2006, at 17:56, Pancho Cole wrote: I use Pro FTP to chroot users to their home directories. see http://www.proftpd.org/ Yes, but the point is they also need to access another directory, owned by them, but well outside of that chroot, all under one login. Not using pro-ftpd, I can't allow ftp access in a chroot to all the files a user needs. Gaby -- Junkets for bunterish lickspittles since 1998! http://www.playr.co.uk/sudoku/ http://weblog.vanhegan.net/
Re: New server
On Sat, May 20, 2006 at 03:45:35PM +0100, Gaby vanhegan wrote: On 20 May 2006, at 15:15, Joachim Schipper wrote: Something's got to give here. I suspect that I'm going to have to un- chroot the ftp daemon. Is there an ftpd somewhere that can prevent users from looking at certain directories? For example, I would like to limit access only to /home/username and /var/www/home/username in ftpd, and prevent access to places like /etc, /usr/local, and so on. A lot of FTP daemons can do that, but I don't really see the point. The protections they offer might or might not be circumventable, but nothing interesting should be readable anyway. If the ftpd runs as the UID of the person that's logged in, they won't be able to access the files they don't own anyway (contents of / etc, and others). But if possible, I'd just like to hide them from view, so they can't even be read. For example, # ls -lFa /etc | grep passwd -rw--- 1 root wheel 2688 May 19 21:57 master.passwd -rw-r--r-- 1 root wheel 2235 May 19 21:57 passwd Would still result in somebody with FTP access being able to download a list of users on the system. I would like to prevent them from doing that if possible. Okay, that's a point. I don't think it matters that much, but... How about the following setup: - All users who should have ftp access are members of the login group ftp-users, and have a home directory /var/ftp (not _in_; /var/ftp is their home directory). This login group includes appropriate values for auth-ftp, and sets ftp-chroot. - This directory is owned by root:wheel and mode 755; equivalent permissions are set for the rest. This is a not a problem, since the home directory need not actually be distinct for each user, as they do not have shell access anyway. If you do need distinct home directories, this is quite possible - use /var/ftp/john. - If it is desired that people have FTP storage that cannot be reached via HTTP, place this under /var/ftp/john (i.e. their home directory, if they have one - see above) - If it is desired that people have FTP storage that cannot be reached via HTTP, but can be reached via anonymous FTP, place this under /var/ftp/pub/john. Obviously, this directory should be created at login. - Place web pages under /var/ftp/users/john - Copy the Apache configuration under /var/ftp, make sure it is only readable by the appropriate user, and Apache cannot read more than necessary. This likely means making the /var/ftp/users/* directories sgid www. - Symlink /var/www to /var/ftp - Fix any and all errors caused by this setup It is worth noting that there's nothing special about /var/ftp; /var/www or /home would do as well. Now, as long as nobody manages to compromise the small bit of the FTP daemon running as root, this should be pretty secure. The stock FTP daemon does not offer any way to get both non-numeric user/group names and a non-user-readable /etc/passwd; other FTPds might. Anyway, ISTR that ProFTPd could do that; I'm quite certain neither stock ftpd nor vsftpd can. I hear that the security record of ProFTPd is not stellar, to say the least. I'm fairly sure that the stock ftpd can't, and I can't find anything in pure-ftpd about it either. That's true, and it led me to comment that its notion of what people can and cannot read might or might not actually work. ;-) Joachim
Re: ksh: typeset screwing up subsequent parameter's array indices?
On Sat, May 20, 2006 at 11:28:26AM -0400, jared r r spiegel wrote: i have half a mind that says otto@/kili@/ckuthe will come down and cluestick me about why what i am seeing is expected behaviour, but the other half expects this to be not what should be happening. the more i play with this the more it seems like i'm just not supposed to be using arrays with integers... :/ (note weird echo output, but the indices are OK ) == $ typeset -i F=0 $ F[1]=1 $ F[2]=3 $ F[93]=29389238 $ echo ${F[*]} 8 8 8 29389238 $ set | grep ^F F[0]=0 F[1]=1 F[2]=3 F[93]=29389238 == and in this one, i'm working on some little dinky functions. the _d[0] parameter ,, the '0' is different every time i run it in a new shell function _dectohex { typeset -i16 _h='' typeset -i10 _d='' while [[ $* != ]]; do _d[++x]=$1 shift done set | grep _d echo [EMAIL PROTECTED] } _dectohex 99 in the above, if you get rid of the typeset -i16 line, it seems to become sane. i'd like to be sane again too :P -- jared [ openbsd 3.9-current GENERIC ( may 1 ) // i386 ]
Re: LUNA: Re: FC5 hogging RAM .... -- Nautilus is a Gecko basednightmare
Is hogging a technical term? I'm not sure that wikipedia has the right definition. Thanks -Original Message- From: Bryan J. Smith [mailto:[EMAIL PROTECTED] Sent: Saturday, May 20, 2006 11:43 AM To: Chris Adams Cc: Linux Group HuntsVegas Subject: Re: LUNA: Re: FC5 hogging RAM -- Nautilus is a Gecko basednightmare On Sat, 2006-05-20 at 11:01 -0500, Chris Adams wrote: Gecko != Mozilla. You have been saying that somehow Nautilus and Mozilla are the same. That's like saying cp = ls because they both use libc. Huh? No, completely different. However, I can't see that Nautilus is one of them. I'm looking at a running copy of nautilus (checking the libraries it has mapped in), the RPM dependencies, and at the source (both the SRPM spec file requirements and the source itself) and I see no reference to Gecko. It has the _old_ GRE. It's not even statically linked, it's _embedded_. That's why you don't see it. Now, maybe Nautilus internalizes Gecko, but it sure doesn't look that way. Oh, a quick Google on 'nautilus Gecko based nightmare' finds one source for that: a NewsForge message board with that exact phrase, followed by another: Indeed!-FUDDING with Misinformation. Oh, what about the Gnome users? Well, their screwed because Nautilus is also a Gecko based nightmare but, being Gnome users, their use to being screwed. We're use to being screwed by those who FUD with misinformation. Nautilus uses the gtkhtml component to display HTML, not gecko. That person is not aware of the Nautilus development. At first, they were using GTKHTML[2] _just_ for browsing. But they are using Gecko for _other_ components as well. -- Bryan J. SmithProfessional, technical annoyance mailto:[EMAIL PROTECTED] http://thebs413.blogspot.com --- Americans don't get upset because citizens in some foreign nations can burn the American flag -- Americans get upset because citizens in those same nations can't burn their own LUNA-LIST help: [EMAIL PROTECTED] To unsubscribe: [EMAIL PROTECTED] To email the list keeper:[EMAIL PROTECTED] LUNA-LIST Web Site: URL:http://luna.huntsville.al.us
Re: Splitting xbaseXY.tgz - stupid idea?
On Sat, May 20, 2006 at 10:09:15AM +0300, Liviu Daia wrote: I have a simpler question: is there any plan to make installing xbase a requirement in the foreseeable future? no. nothing in {base,comp,man,misc,game,etc}XX.tgz depends on anything from xbaseXX.tgz, and that is extremely unlikely to ever change. So what you're saying here is that installing 30MB of xbase without the user requesting it is acceptable, if you're installing a port that depends on something in xbase, then you are requesting xbase. -- [EMAIL PROTECTED]
Re: ksh: typeset screwing up subsequent parameter's array indices?
On Sat, 20 May 2006, jared r r spiegel wrote: On Sat, May 20, 2006 at 11:28:26AM -0400, jared r r spiegel wrote: i have half a mind that says otto@/kili@/ckuthe will come down and cluestick me about why what i am seeing is expected behaviour, but the other half expects this to be not what should be happening. this looks like a bug, -Otto the more i play with this the more it seems like i'm just not supposed to be using arrays with integers... :/ (note weird echo output, but the indices are OK ) == $ typeset -i F=0 $ F[1]=1 $ F[2]=3 $ F[93]=29389238 $ echo ${F[*]} 8 8 8 29389238 $ set | grep ^F F[0]=0 F[1]=1 F[2]=3 F[93]=29389238 == and in this one, i'm working on some little dinky functions. the _d[0] parameter ,, the '0' is different every time i run it in a new shell function _dectohex { typeset -i16 _h='' typeset -i10 _d='' while [[ $* != ]]; do _d[++x]=$1 shift done set | grep _d echo [EMAIL PROTECTED] } _dectohex 99 in the above, if you get rid of the typeset -i16 line, it seems to become sane. i'd like to be sane again too :P -- jared [ openbsd 3.9-current GENERIC ( may 1 ) // i386 ]
Re: ServeRAID 4M
On Sun, May 21, 2006 at 12:39:30AM +0600, Anton Maksimenkov wrote: Hi. Reading through misc archives I found this post http://marc.theaimsgroup.com/?l=openbsd-miscm=112454454105020w=2 Currently I got number of IBM ServeRAID 4M controllers (plus batteries packs) in couple with some IBM machines (eServer 6BY and some other models). It's dramatic for me because of some ServeRaid cards are supported and some are not. And of course I have to to ask - is the ServeRAID 4M supported (some how)? When I tried to install 3.9 on it the installer ends with no disk found... But the card is at least recognized as (full dmesg below): IBM ServeRAID rev 0x00 at pci1 dev 7 function 0 not configured I want to know what does mean that message? Is the card MAY work (after some)? May be some BIOS, card BIOS or UKC options must be carefully set? 'not configured' typically means the kernel knows what it is, but doesn't know what to do with it. I searched through /usr/src/sys/dev/pci/pcidevs and found some claws for ServeRAID, but not found ServeRAID elsewhere... I tried to reproduce my searches with ami (MegaRAID - I added that card to that machine for experimental reasons) and I found some - man page, some claws in src. So where the definitions for ServeRAID? I respect developers anger about IBM behaviour... But if job for these controllers was already DONE and it might work somehow - so can I use it? I'm not a developer or a hardware guru, so I'll not comment on the likelihood of it being supported 'soon' (for values of soon in 'it's in -current', 'undergoing testing', 'partially done', 'someone is working on it', 'someone is interested', 'it would be neat to have'). Joachim
Re: LSI MegaRaid non-hotspare
Depends on which bug you hit. If the BIOS shows it is a hotspare you're golden. If it does not show it as a hotspare you want to upgrade (or use the bios to create the hotspare). Gaby vanhegan wrote: On 20 May 2006, at 16:28, Marco Peereboom wrote: I fixed this in current. You can simply just upgrade the ami files to -current and build a 3.9 that is mostly RELEASE. Was it a functional problem or just a cosmetic one? If I leave it as it is, is it going to cause any real problems for me? Gaby -- Junkets for bunterish lickspittles since 1998! http://www.playr.co.uk/sudoku/ http://weblog.vanhegan.net/
Re: ksh: typeset screwing up subsequent parameter's array indices?
On Sat, May 20, 2006 at 08:38:38PM +0200, Otto Moerbeek wrote: [typeset -i F and ${F[*]} weirdness] this looks like a bug, It's caused by the fact that for integers str_val() returns the address of a static buffer and that the loop over the array elements in varsub() just copies the pointer returned by str_val() into a pointer vector (eval.c, line 793). Would strdup()ing and later free()ing all the strings be an option? Or do you consider that overkill? Ciao, Kili
Re: ksh: typeset screwing up subsequent parameter's array indices?
On Sat, 20 May 2006, Otto Moerbeek wrote: On Sat, 20 May 2006, jared r r spiegel wrote: On Sat, May 20, 2006 at 11:28:26AM -0400, jared r r spiegel wrote: i have half a mind that says otto@/kili@/ckuthe will come down and cluestick me about why what i am seeing is expected behaviour, but the other half expects this to be not what should be happening. this looks like a bug, Actually, it looks like two bugs. One is typeset -i related, the other occurs if you do: [EMAIL PROTECTED]:1]$ A=a [EMAIL PROTECTED]:2]$ A[1]=b [EMAIL PROTECTED]:3]$ set A[593830152]=a A[1]=b ... The diff below seems to fix the latter; I'm still investigating the typeset -i stuff. -Otto Index: var.c === RCS file: /cvs/src/bin/ksh/var.c,v retrieving revision 1.29 diff -u -p -r1.29 var.c --- var.c 13 Mar 2006 08:21:37 - 1.29 +++ var.c 20 May 2006 20:29:02 - @@ -1107,12 +1107,10 @@ arraysearch(struct tbl *vp, int val) size_t namelen = strlen(vp-name) + 1; vp-flag |= ARRAY|DEFINED; - + vp-index = 0; /* The table entry is always [0] */ - if (val == 0) { - vp-index = 0; + if (val == 0) return vp; - } prev = vp; curr = vp-u.array; while (curr curr-index val) {
Re: LSI MegaRaid non-hotspare
* Marco Peereboom [EMAIL PROTECTED] [2006-05-20 21:48]: Depends on which bug you hit. If the BIOS shows it is a hotspare you're golden. If it does not show it as a hotspare you want to upgrade (or use the bios to create the hotspare). no, that is incorrect. even if the bios does hsow you teh drive as hotspare, it might not be used as one; you have to to put it back to unused using the bios and then mark as hotspare in the bios. that is the only 100% reliable way for the moment, unfortunately. -- BS Web Services, http://www.bsws.de/ OpenBSD-based Webhosting, Mail Services, Managed Servers, ... Unix is very simple, but it takes a genius to understand the simplicity. (Dennis Ritchie)
Re: Sound card with supported digital out
On Sat, May 20, 2006 at 05:46:42AM +0200, Jan Johansson wrote: andrew fresh [EMAIL PROTECTED] wrote: Is there a supported sound card that supports digital outputs? I think your best bet is USB audio. I have a simple USB audio stick that does optic digital signal or headphones under OpenBSD. I have tried one of those, I had forgotten about that. The problem with the USB digital output that I have tried is that it does not do AC3/DTS passthrough, all it does is output 2 channel PCM over the optical digital connection. I believe the one I tried was a Turtle Beach Audio Advantage Micro. If there is USB audio that will do AC3/DTS passthrough on OpenBSD, I would be happy with that. l8rZ, -- andrew - ICQ# 253198 - JID: [EMAIL PROTECTED] BOFH excuse of the day: wrong polarity of neutron flow
Re: ksh: typeset screwing up subsequent parameter's array indices?
On Sat, 20 May 2006, Matthias Kilian wrote: On Sat, May 20, 2006 at 08:38:38PM +0200, Otto Moerbeek wrote: [typeset -i F and ${F[*]} weirdness] this looks like a bug, It's caused by the fact that for integers str_val() returns the address of a static buffer and that the loop over the array elements in varsub() just copies the pointer returned by str_val() into a pointer vector (eval.c, line 793). Would strdup()ing and later free()ing all the strings be an option? Or do you consider that overkill? Indeed, the static buffer is the problem. A dynamically alloc'ed string could work, but the problem would be where/when to free it. In the case of formatstr an allocated string is returned, so we have already a mem leak here, it seems. I remember seeing a NetBSD commit related t0 formatstr handling: http://cvsweb.netbsd.org/bsdweb.cgi/src/bin/ksh/var.c.diff?r1=1.12r2=1.13 But I must sleep now... -Otto
Re: ServeRAID 4M
Joachim Schipper wrote: 'not configured' typically means the kernel knows what it is, but doesn't know what to do with it. More specifically, it means that the kernel knows the PCI device's ID and vendor, but doesn't have a driver to hook it to. FreeBSD supports it with the ips driver and it appears to be non-BLOB; the original commit message also implies that it's actually an Adaptec card, so it might be as easy as hooking it to an existing Adaptec driver. (I doubt it, but it may be worth a shot.)
Re: ksh: typeset screwing up subsequent parameter's array indices?
On Sat, 20 May 2006, Otto Moerbeek wrote: On Sat, 20 May 2006, Matthias Kilian wrote: On Sat, May 20, 2006 at 08:38:38PM +0200, Otto Moerbeek wrote: [typeset -i F and ${F[*]} weirdness] this looks like a bug, It's caused by the fact that for integers str_val() returns the address of a static buffer and that the loop over the array elements in varsub() just copies the pointer returned by str_val() into a pointer vector (eval.c, line 793). Would strdup()ing and later free()ing all the strings be an option? Or do you consider that overkill? Indeed, the static buffer is the problem. A dynamically alloc'ed string could work, but the problem would be where/when to free it. In the case of formatstr an allocated string is returned, so we have already a mem leak here, it seems. I remember seeing a NetBSD commit related t0 formatstr handling: http://cvsweb.netbsd.org/bsdweb.cgi/src/bin/ksh/var.c.diff?r1=1.12r2=1.13 And http://cvsweb.netbsd.org/bsdweb.cgi/src/bin/ksh/var.c.diff?r1=1.13r2=1.14f=u actually solves that mem leak But I must sleep now... -Otto
Re: ksh: typeset screwing up subsequent parameter's array indices?
How about this? If I see things correctly, the ATEMP allocation should be cleaned up automatically Running a little test loop does not show a leak. Both bugs are fixed, and array entries are nice integer vals. [EMAIL PROTECTED]:189]$ cat t typeset -i F=0 F[1]=1 F[2]=3 F[93]=29389238 F[98]=444 F[100]=a echo ${F[*]} typeset -i [EMAIL PROTECTED]:190]$ obj/ksh t 0 1 3 29389238 444 0 F[0]=0 F[1]=1 F[2]=3 F[93]=29389238 F[98]=444 F[100]=0 MAILCHECK=600 OPTIND=1 PPID=9226 RANDOM SECONDS=0 TMOUT=0 [EMAIL PROTECTED]:191]$ -Otto Index: var.c === RCS file: /cvs/src/bin/ksh/var.c,v retrieving revision 1.29 diff -u -p -r1.29 var.c --- var.c 13 Mar 2006 08:21:37 - 1.29 +++ var.c 20 May 2006 21:54:12 - @@ -293,7 +293,7 @@ str_val(struct tbl *vp) else { /* integer source */ /* worst case number length is when base=2, so use BITS(long) */ /* minus base # numbernull */ - static char strbuf[1 + 2 + 1 + BITS(long) + 1]; + char strbuf[1 + 2 + 1 + BITS(long) + 1]; const char *digits = (vp-flag UCASEV_AL) ? 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ : 0123456789abcdefghijklmnopqrstuvwxyz; @@ -322,6 +322,8 @@ str_val(struct tbl *vp) *--s = '-'; if (vp-flag (RJUST|LJUST)) /* case already dealt with */ s = formatstr(vp, s); + else + s = str_save(s, ATEMP); } return s; } @@ -1107,12 +1109,10 @@ arraysearch(struct tbl *vp, int val) size_t namelen = strlen(vp-name) + 1; vp-flag |= ARRAY|DEFINED; - + vp-index = 0; /* The table entry is always [0] */ - if (val == 0) { - vp-index = 0; + if (val == 0) return vp; - } prev = vp; curr = vp-u.array; while (curr curr-index val) {
Re: xmms does not run smoothly
Thanks for the replies so far :) Sorry for not replying faster, but here goes: Emmanuel Jarri wrote: The workaround I use is to increase buffer size to its maximum, i.e. 13MB, with 50% to upper pre buffer. It works quite nicely, but I feel it's a dirty workaround... My xmms will not go higher than 4096 kb and 90% pre-buffering, and, unfortunately, those values do not make the temporary freezes go away. If it worked for me I would also think of it as a dirty workaround :) Ted Unangst wrote: the two solutions are to prescroll the entire playlist (slowly, so there are no gaps) or to switch to librthread (which is not done, but worked for xmms before anything else). if you haven't heard of librthread, then i don't think it'd be good to switch, but the problem is being worked on. Actually, I already use the option read info on load, so I do not experience freezes when scrolling my playlist. However, the freezes appear frequently anyway, e.g. when xmms opens a dialog that reads directory information from the disk, and therefore still annoys me. I suspect my version of the problem is a bit different from what other people report, since the execution of heavy programs, such as Mozilla Firefox and Thunderbird, also disturbs xmms and causes short lags in the sound. I have not heard about librthread and would rather like to try another player than hacking xmms to get smoother sound on my system. Can you tell me more about what is being worked on and by who? Doug Clements wrote: Check this: http://www.geocities.com/phileosophos/tech/pcilatency.html Thanks for the link - I found it educational. It presents PowerStrip as the solution, however, as you might have guessed by now, I am not running Microsoft Windows :) It makes me wonder how to change the PCI priorities in OpenBSD... hmm... I will try to look into that. If somebody knows that this is a wrong path to follow, then please tell me. /Martin
Re: xmms does not run smoothly
On 2006/05/21 01:00, Martin Toft wrote: I have not heard about librthread and would rather like to try another player than hacking xmms to get smoother sound on my system. Can you tell me more about what is being worked on and by who? You can read about rthreads here, http://www.openbsd.org/papers/eurobsd2005/tedu-rthreads.pdf Doug Clements wrote: Check this: http://www.geocities.com/phileosophos/tech/pcilatency.html Thanks for the link - I found it educational. It presents PowerStrip as the solution, however, as you might have guessed by now, I am not running Microsoft Windows :) It makes me wonder how to change the PCI priorities in OpenBSD... hmm... I will try to look into that. If somebody knows that this is a wrong path to follow, then please tell me. Some motherboards (e.g. some VIA-based socket7) have problems (including crackly/stuttering sound), fiddling with PCI priorities is a possible work- around.
Re: an easy way to black list IP's
* Darrin Chandler [EMAIL PROTECTED] [060520 10:21]: On Sat, May 20, 2006 at 09:49:31AM -0400, Jim Razmus wrote: Take a look at mail/relaydb in the ports tree. Also check the archives as this has been discussed at depth and included several solutions. Why ports instead of packages? Notably since he's already tried relaydb (and it doesn't do forwarded messages). Are there other flavours of interest? -- Darrin Chandler| Phoenix BSD Users Group [EMAIL PROTECTED] | http://bsd.phoenix.az.us/ http://www.stilyagin.com/ | LOL. That's what I get for scanning, not _reading_, the original post. Sorry for the noise. Man am I embarrassed. Jim
Re: an easy way to black list IP's
Hey I got exactly what you are looking for, its pretty easy. You need relaydb and procmail. Setup a user called 'spam' then in /home/spam/ ... # cat .forward |/home/spam/procspam.sh # cat .procmailrc # .procmailrc ORGMAIL=/var/mail/$LOGNAME PATH=/usr/bin:/usr/local/bin MAILDIR=$HOME/.mailspool # all mailboxes are in .mailspool/ #DEFAULT=$HOME/.mailspool/spam LOGFILE=/dev/null SHELL=/bin/sh :0b: spam # cat procspam.sh #!/bin/sh HOME=/home/spam PATH=/sbin:/usr/sbin:/bin:/usr/bin:/usr/X11R6/bin:/usr/local/sbin:/usr/local/bin /usr/local/bin/procmail relaydb -f /var/spamd/.relaydb -i /var/spamd/whitelist.relaydb | cat spam | grep -A 1000 Received: | relaydb -bf /var/spamd/.relaydb rm $HOME/spam and then of course spamd.conf .. relaydb-black:\ :black:\ :msg=SPAM. Your address %A is in my relaydb list.:\ :method=exec:\ :file=/usr/local/bin/relaydb -4lb -f /var/spamd/.relaydb: Craig Hammond wrote: Hi all, I'm using spamd and it does a great job. What I'm trying to figure out is how to easily add the IP's of the sending mail server for the few spam that still get through. By easy, I mean for clients of mine who use Exchange/Outlook, where I put a obsd box running spamd in front of Exchange. I am trying to find a way where I could tell my clients that when some spam does get through, just forward that spam to a particular email address. Some process will extract the IP of the MTA that sent the spam and blacklist it. I installed and played around with relaydb from ports, but that doesn't work with emails that have been forwarded. Any ideas??
software load balancing
I worked with a customer once that had a software based load balancing solution. I liked the way it worked. While I was working on the box if I was going to take the service down for maintenance I could tell the local agent and the box was removed from the pool of servers. Anyone know of something like this that runs on OpenBSD? The master controller part on openbsd would be great, with agents for various other operating systems as well. carp/pfsync is great, but I'm thinking of a times when the application that needs to be load balanced won't run on openbsd, say only on Solaris. Thanks, Chad
Re: an easy way to black list IP's
You do know that headers can be forged right? So an automagic forward - |/script - blacklist from a pissed off user can end up blacklisting a legitimate MTA. This is a good point, if you use the scripts I sent you may want to modify them to look for a password, should be simple enough. Mike Spenard
Re: [patch] Intel 945G/GM AGP support (including 945GM for X.org)
vladas wrote: making all in programs/Xserver/Xext/extmod... make: don't know how to make /usr/include/stdarg.h. Stop in /usr/Xbld/xc/program s/Xserver/Xext/extmod. *** Error code 2 You probably don't have the compiler installed at all. Did you install comp39.tgz? If not, see section 4.10 of the FAQ. Btw, there should be snapshots coming up now which already include the 945 patches for both the kernel and X.org, so maybe it's easier to try one of those. Cheers, Dimitry
Re: ksh: typeset screwing up subsequent parameter's array indices?
On Sat, May 20, 2006 at 11:59:13PM +0200, Otto Moerbeek wrote: How about this? If I see things correctly, the ATEMP allocation should be cleaned up automatically Running a little test loop does not show a leak. Both bugs are fixed, and array entries are nice integer vals. ... Index: var.c === RCS file: /cvs/src/bin/ksh/var.c,v retrieving revision 1.29 diff -u -p -r1.29 var.c ... tested that one against the 'echo ${thing[*]}' weirdness - fixes that right up, afaict jared