OpenBSD Brasil
Configuração de um servidor NTP no OpenBSD 5.0

Procedimento

  1. No arquivo /etc/ntpd.conf descomente a linha listen * e no lugar do * coloque o IP da interface da rede local, ficando assim:

    listen 192.168.0.1

  2. No arquivo /etc/rc.conf.local coloque a linha:

    ntpd_flags=”-s”

  3. Inicie o serviço executando:

    /etc/rc.d/ntpd start

guilhermedeoliveirachagas:

PessoALL, como um bom novato tenho a obrigação de documentar a compartilhar os conhecimentos adquiridos nesse início da minha jornada com o OpenBSD. Hoje irei compartilhar com vocês uma das maneiras de se instalar OpenBSD 5.0 e Windows 7 Ultimate configurando-os como Dual Boot.

1. Instalação e…

Aplicando a 1a atualização do OpenBSD 5.0

Saiu uma atualização para o Bind. Como proceder se você possui o OpenBSD?

O Bind faz parte da árvore do OpenBSD. Isto significa que o Bind possui patches feitos pela equipe de desenvolvedores para torná-lo mais seguro/estável. Portanto, para fazer atualização dele, basta seguir as instruções do patch de atualização. Por exemplo, se você usa a versão 5.0 do OpenBSD, você já tem 1(uma) atualização pra fazer.

Organizando os fontes da árvore do OpenBSD

Para tornar seu OpenBSD estável (stable) Execute o seguinte procedimento,como root: cd /root ftp -V ftp://ftp.openbsd.org/pub/OpenBSD/5.0/sys.tar.gz ftp -V ftp://ftp.openbsd.org/pub/OpenBSD/5.0/src.tar.gz cd /usr/src tar -zxvf /root/sys.tar.gz tar -zxvf /root/src.tar.gz

Fazendo a atualização do Bind

Quando um patch de segurança para a ávore do OpenBSD é corrigido, o mesmo é divulgado imediatamente na lista security-announce, algumas horas depois aparece no site.

Você encontrará uma correção. Baixe o arquivo para o host OpenBSD usando , como root, os comandos:

cd /root
ftp -V http://ftp.openbsd.org/pub/OpenBSD/patches/5.0/common/001_bind.patch"

E siga a instruções contidas no arquivo:

Apply by doing:
    cd /usr/src
    patch -p0 < /root/001_bind.patch

And then rebuild and install bind:
    cd usr.sbin/bind
    make -f Makefile.bsd-wrapper obj
    make -f Makefile.bsd-wrapper depend
    make -f Makefile.bsd-wrapper
michael:


Building VPN’s with OpenBSD and IPSEC

0xac0:

Data: 2011-09-15 14:54:00 -03:00

Preâmbulo

Definitivamente, estive em apuros por causa duma limitação do gerenciador de boot do OpenBSD. A falha consistia em travar o computador no momento do boot do CD, no momento em que ele tenta reconhecer a memória RAM do computador.

Local do arquivo

Ao…

0xac0:

Data: 2011-09-13 13:12:00 -03:00

Introdução

É comum usuários de outros sistemas operacionais, principalmente os Unix-Like, quando começam usar o OpenBSD se perguntarem como se faz a para deixar o sistema atualizado. -Como se dá o “apt-get upgrade”? -Eles perguntam. Isto em geral ocorre pelo fato…

firemyst:

this aims for a very lean install of php and mongo on openbsd (dated 11-2011)

so real quick without pear, pecl and unnecessary garbage -install or upgrade to your preferred openbsd flavor (stable, release, current)

 $ sudo pkg_add -vi php $ sudo pkg_add -vi mongodb

this should get you php 5.3x…

O que há de novo no OpenBSD 5.0

Com data prevista de lançamento previsto para 1 de Novembro de 2011, a nova versão traz consigo mais de 770 alterações, entre novas funcionalidades e correções.

Segue abaixo algumas:

Drivers

  • Added dfs(4/MACPPC) driver to support the Dynamic Frequency Switching feature found on some laptops.
  • Support additional L2C variants and L1D (AR813x/AR815x chips) in alc(4).
  • Update to sendmail(8) 8.14.5.
  • Various drivers have been adjusted to use PCI Message Signaled Interrupts on amd64, i386, macppc and sparc64.
  • New AMD K10/K11 pstate driver allows setperf and apm to change CPU frequences on newer AMD CPUs.
  • Unified various macppc gpio(4) access methods that take an offset relative to the mac-io bus base address. Needed for upcoming dfs(4/MACPPC) support.
  • Fixed arguments to arm pmap(9) poolinit: alignment of alignment of L2TABLESIZEREAL is at offset 0 and not at offset L2TABLESIZE_REAL.
  • Make pci(4) pass flags down the PCI bus hierarchy.
  • Added pci(4) register definitions for PCI MSI capability.
  • Stop leaking swapslots in uvm(9) when doing a uvmkmpgremove and a page is in swap only.
  • Removed uvm(9) vmpagelookup_freelist().
  • Implemented correct prologue and epilogue for hppa64 machine-dependent init.
  • Make mips common kernel code set octeon’s internal counter clock speed to its processor clock.
  • Fixed aucat(1) option handling and enable TCP in midicat(1).
  • Refactored queue allocation and initialization into wdc(4) wdcallocqueue() function, and let attachment code call this rather than malloc(9). This prevents re-initialization of the queue in shared queue chipsets.
  • Initialize the wdc(4) atadrivedatas structures earlier in wdcattach() so that chip-specific drv_probe routines can assume they’ve already been initialized.
  • Make sndio(7) siopsleep() use an array of SIOMAXNFDS pollfd structures rather than a single one.
  • Initialize the ‘pstate’ field of the aucat(1) wav structure.
  • Fixed a few minor issues in i386 hibernate support code relating to improper swap device determination and memory range calculation.
  • Fix memory handling in octeon machine-dependent code.
  • Eliminated a few unused wdc(4) capability flags (WDCCAPABILITYHWLOCK, WDCCAPABILITYATANOSTREAM, and WDCCAPABILITYATAPINOSTREAM).
  • Simplified physio(9) thanks to the fact that buffers now come out of a pool rather than a global list of statically allocated structures and aren’t shared.
  • Removed uvm_pglist.h from the tree.
  • Make ubsec(4) interrupt routine acknowledge only the interrupts it can process.
  • Make sure amd64 AES session id checks look at the lower 32 bits of crp_sid.
  • Make hppa64 gateway page accessible to all userland processes.
  • Reverted atapiscsi(4) to only attempting on IDENTIFY command against directly attached devices as in the pre-port-multiplier code.
  • Make the “mute” key work on macppc keyboards.
  • Make urndis(4) attach to Samsung Galaxy S.
  • Make sure hppa64 restore sr4 at the very end of locore.S to avoid further loads from the trapframe to be done at the wrong address space.
  • Make scsi(4) skip leading blanks and collapse multiple white spaces into when when printing scsi device ids.

X11

  • Fixed client and group cycle defines in cwm(1).
  • Fixed xf86-input-ws on xserver 1.9 and earlier.
  • Enabled xf86-input-synaptics on i386 and amd64.
  • Added support to new wscons(4) ioctl WSMOUSEIO_SETMODE in xf86-input-synaptics.
  • Updated xf86-input-acecad to version 1.5.0, xf86-video-chips to 1.2.4, xf86-video-vmware to 11.0.3, xf86-video-siliconmotion to 1.7.5, xlsclients to 1.1.2.
  • On cwm(1) map, prevent it from warping the windows that are marked as ignored.
  • Put back cwm(1) window resize back to 60 Hz.
  • Make cwm(1) menu window aware of xinerama(3) info.
  • Keep synaptics touchpad to the current wscons(4) behaviour until WSMOUSEIO_SETMODE ioctl is issued to switch to synaptics mode.

Rede

  • Change ‘set skip on <…>’ to work with interface groups.
  • Disable pipex for L2TP on disconnect.
  • Pre-allocate memory in ipsec(4) package to avoid sleeping after performing a lookup, which may lead to a race.
  • Make pfsync(4) use timeout(9) timeout_del return value to check if the timeout is actually removed when undeferring a packet.
  • Make tftp-proxy(8) use ‘divert-to’.

Comandos

  • In addition to relative resizing, allow absolute resizing of partitions in auto-allocated labels with disklabel(8).
  • Allow specifying k/m/g/… suffixes in newfs(8) -S and -s options.
  • Fixed some warnings in adduser(8).
  • Make ssh(1) use FD_CLOEXEC consistently.
  • Cleaned up adduser(8) handling of email messages.
  • Make gdb(1) handle lazy relocation stubs as Linux does.
  • Make tmux(1) reset last pane on break-pane. Fixes a problem reported in Debian bug #622677.
  • Make tmux(1) reset last pane on swap-pane across windows. Fixes a crash.
  • Prevent tmux(1) from dragging on click, only select.
  • Make ssh(1) warn on unexpected key type in keyparseprivate_type().
  • Removed support for authorized_keys2, a relic from the early days of protocol V2, in ssh(1).
  • Make dhclient(8) more friendly to sequential option processing by always starting DHCP packet options with DHODHCPMESSAGE_TYPE. Improved working with Nortel NetIP DHCP server.
  • Fixed memory leaks in tmux(1) command capture pane.
  • Fixed a memory leak in tmux(1) commands if cmdpanesession succeed.
  • Added a new option to tmux(1), mouse-resize-pane which, when on, allows panes to be resized by dragging their borders.
  • Make tmux(1) use the tsl and fsl terminfo(5) capabilities to update terminal title and automatically fill them in on terminals with the XT capability.
  • Added a small memory optimization in fsck_ffs(8).
  • Improved tmux(1) behaviour when TTY allocation fails: if RequestTTY is set to ‘auto’ make it not treat a TTY allocation error as fatal and just restore the local TTY.
  • Make more silent ssh(1) debug() logs by detecting that it’s trying to load a private key in keytryload_public() and returning early.
  • Make sure sysmerge(8) handle first /etc/group and /etp/master.passwd in case it need to install files or directories with newly added user/group ownerships.
  • Introduced a ‘freeze’ flag in tmux(1) which make it ignore any move or resize requests made on the windows it’s applied.
  • Added a RequestTTY ssh_config(5) option to allow configuration-based control over tty(4) allocation, like ssh [-tT].
  • Make ssh_config(5) support negated host matching.
  • Added a %L expansion (short-form of the local host name) for ssh(1) ControlPath.
  • Set ssh(1) traffic class for IPv6 traffic as it’s done for IPv4 TOS. Fixes bz#1855.
  • Make sysmerge(8) create the directory holding the link it’s about to create if it does not exist.
  • Fixed memory leak in ssh(1). Fixes bz#1849.
  • Added disklabel(8) support in tunefs(8).
  • Make ssh(1) gracefully fall back when ControlPath is too large for a sockaddr_un.
  • Allow ssh-add(1) to read key from standard input with ssh-add - .
  • Make atactl(8) capable of reading disklabel(8) UIDs.
  • Make tmux(1) change window with mouse wheel over status line if mouse-select-window is on.
  • Prevent use of strnvis(3) in tmux(1) title as it breaks UTF-8.
  • Make tmux(1) check if mouse-select-pane is on, not off when setting mouse flags.

Daemons

  • Fixed reload support in relayd(8).
  • Fixed type warnings reported by clang in smtpd(8).
  • Fixed wrong id for UDPENCAPTRANSPORT_DRAFT in isakmpd(8).
  • Fixed an off-by-one that made smtpd(8) skip an “invalid” bucket that was actually valid.
  • Make relayd(8) use the proc.c privsep API/commodity functions based on work for iked(8) and smtpd(8).
  • Fixed segfault in smtpd(8) newaliases after global env move.
  • Prevent security(8) from complaining about a group(5) line with a single “+” as “wrong number of fields”, that abbreviated syntax for NIS map of groups is explicitly allowed by group(5). Warn if this isn’t the last line of group(5) though.
  • Started a work in ospfd(8) to support opaque LSA.
  • Force loopback interfaces to IFSTALOOPBACK in ospf6d(8).
  • Updated relayd(8) logging and debug functions to use the C99 func macro instead of static function names.
  • Allow a user to specify relayd(8) root priority.
  • Fixed check of errors in bgpd(8) sessions.
  • Make iscsid(8) handle logins more correctly.
  • Prevent smtpd(8) from fork-bombing on startup when there are lots of mails in the offline queue by using a wait list to keep the number of forked processes below a reasonable limit when enqueueing.

Bibliotecas

  • Make calls to malloc(3) malloc_dump() safer by avoiding file pointer computation for stats.
  • Introduced leak detection code for MALLOC_STATS in malloc(3).
  • Fixed bug in glob(3).
  • Make timeout(9) timeout_del able to tell the caller if it actually did remove a timeout or not.
  • Added a wprintf(3) man pages.
  • Always free the multibyte->wchar conversion buffer allocated in vfwprintf(3) __mbsconv().
  • Make malloc(3) start scanning the bits of the chunk at a random position to take the first available free slots instead of starting from position zero and skipping a random number of free slots. Make things faster.