hey so i've upgraded to a locally built 2.6.14.4 ... and tail -f is still
broken with that kernel. i'm wondering now what's different between
debian 2.6.x kernels and my locally built one (it's on a production server
though so i can't experiment too much).
anyhow i'm really surprised nobody
aha!
i figured out what's going on... it's not a 2.4 vs. 2.6 kernel
difference... the difference is that on one of my hosts all of my log
files have the append only attribute.
check this out:
/tmp# dpkg -s coreutils
...
Version: 5.93-5
/tmp# touch append_only
/tmp# tail -f append_only
/tmp#
Package: initrd-tools
Version: 0.1.84
currently for a root on md raid mkinitrd does something like this:
mdadm -A /dev/md3 -R -u 2b3a5b77:c7b4ab81:a2b8322a:db5c4e88 /dev/sdb4 /dev/sda4
however this has the problem that it will require the root raid devices to
always reside at /dev/sda4 and
Package: clamav-freshclam
Version: 0.87-1
a couple times a month i find a freshclam which has been stuck on a read
from fd 4 for a few days... fd 4 is its network socket, and it seems to be
stuck in the middle of a tcp session, probably the other end has
disappeared. there really should be an
Package: spell
Version: 1.0-15
try this:
% echo a b c a
% spell a
it hangs forever...
i dunno, but the source code looks like it has some confusion with pin vs.
pout... the patch below seems to fix it.
-dean
--- spell-1.0/spell.c.dg2005-10-22 11:26:24.0 -0700
+++
Package: coreutils
Version: 5.93-2
you can't use O_NONBLOCK on files in 2.4... and so tail -f is broken
because it apparently requires that now as of 5.93:
tail: /var/log/apache/access_log: cannot change nonblocking mode: Operation not
permitted
i don't think that should be a fatal error (or
ugh this whole blocking optimization stuff in tail_forever just seems
broken... i'm not sure the correct fix, but the following fix seems to do
the job for me.
i've tested doing tail -f and -F on one and multiple files on 2.6 and 2.4
... i've done other nonsense tests like tail -f on stdin,
Package: rtorrent
Version: 0.6.1-1
scripts/checks.m4: ok=`echo ibase=16; if($hex_ver=$check_hex) $hex_ver
else 0 | bc`
needs bc installed... it's there in 0.6.2 upstream as well.
just for pedantry could you add a build-depends: bc?
thanks
-dean
--
To UNSUBSCRIBE, email to [EMAIL
Package: munin-node
Version: 1.2.5-1
for the exact right length of sensors name there's no spaces following the
colon... for example:
% sensors |grep Temp
CPU2_Temp:+39.25 C (low =+0 C, high = +90 C)
CPU1_Temp:+42.00 C (low =+0 C, high = +90 C)
i fixed all three regex even
Package: mdadm
Version: 2.5.5-1
Severity: grave
it's dangerous to generate an mdadm.conf and start running arrays
automatically at install time! i nearly got bit by this.
i marked this grave because there's a potential for data loss with the
current install scripts.
i had 4 disks which i had
Package: mdadm
Version: 2.5.5-1
Severity: grave
even though i have INITRDSTART='none' in my /etc/default/mdadm and rebuilt
the initrd, it still goes and does array discovery at boot time.
this is marked grave because it can cause dataloss if drives with stale
superblocks are put together in an
On Mon, 13 Nov 2006, martin f krafft wrote:
severity 398312 important
tags 398312 unreproducible moreinfo
thanks
even though i have INITRDSTART='none' in my /etc/default/mdadm and rebuilt
the initrd, it still goes and does array discovery at boot time.
piper:/tmp/cdt.d.Ns8889# grep
On Mon, 13 Nov 2006, martin f krafft wrote:
also sprach dean gaudet [EMAIL PROTECTED] [2006.11.13.1107 +0100]:
which causes the is_true() in info() to return 1 which causes the set -e
to terminate the script.
What shell are you using?
my SHELL=/bin/zsh, but that won't affect the script
On Mon, 13 Nov 2006, martin f krafft wrote:
also sprach dean gaudet [EMAIL PROTECTED] [2006.11.13.1116 +0100]:
right, now i know that i should create an /etc/default/mdadm
*before* i install mdadm... because unlike other packages, mdadm
does potentially dangerous things just by installing
Package: initramfs-tools
Version: 0.85a
the run_scripts() function should respect the same naming conventions as
run-parts(8) ... in particular if my editor creates foo~, foo.bak,
.foo.swp files run_scripts() will try to run them. ditto for foo,v.
run_scripts() should also not attempt to
On Mon, 13 Nov 2006, martin f krafft wrote:
severity 398310 important
retitle 398310 let user choose when to start which array
tags 398310 confirmed help
thanks
also sprach dean gaudet [EMAIL PROTECTED] [2006.11.13.0230 +0100]:
i had 4 disks which i had experimented with sw raid10
Package: geoip
Version: 1.3.17-1
the patch below has been accepted into the upstream repository, and will
appear next release... however it may be worth updating the debian package
in the interim.
-dean
- Original message -
From: dean gaudet
Date: Mon, 4 Sep 2006 18:50:28 -0700 (PDT
Package: ddrescue
Version: 1.10-1
there's a new upstream 1.12... which makes O_DIRECT actually work properly
amidst other things. it'd be cool if you could update the debian package.
thanks
-dean
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact
Package: memcached
Version: 1.1.12-1
seems like the safest default install would be to run as a non-root
user... (and i was gonna say listen on 127.0.0.1 but someone beat me to
it).
thanks
-dean
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact
if you apt-get remove bc you get this error during configure of rtorrent
source package:
checking for curl = 7.12.0... ./configure: line 13046: bc: command not
found 7.15.5
now look at the code in scripts/checks.m4:
ver=`curl-config --version | sed -e s/libcurl //g`
On Wed, 20 Sep 2006, Jose Luis Rivas Contreras wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
dean gaudet escribió:
if you apt-get remove bc you get this error during configure of rtorrent
source package:
^^
I'm sorry but I doesn't see this error:
split
paty:/home
... thanks!
-dean
-- Forwarded message --
Date: Wed, 27 Sep 2006 12:15:08 -0700 (PDT)
From: dean gaudet [EMAIL PROTECTED]
To: Stephen Hemminger [EMAIL PROTECTED]
Cc: Jeff Garzik [EMAIL PROTECTED], [EMAIL PROTECTED],
netdev@vger.kernel.org
Subject: Re: mii-tool gigabit support.
On Tue
Package: console-tools
Version: 1:0.2.3dbs-65
the various /etc/console-tools/config settings aren't applied to the vga
device if i boot my system with console=ttyS0,115200 command-line option.
now i know... you're thinking you said the console is serial, not vga...
but then, that begs the
On Sun, 3 Dec 2006, Frans Pop wrote:
On Sunday 03 December 2006 22:34, dean gaudet wrote:
the linux-image .postrm script is (through some mechanism) invoking
/sbin/update-grub.
/sbin/update-grub gives a warning now:
You shouldn't call /sbin/update-grub. Please call /usr/sbin/update
Package: debconf
Version: 1.5.3
apache-1.3.34-2 package can't be configured when debconf 1.5.3 is on the
box... works fine when 1.5.2 is on the box...
tail of set -x output looks like so:
+ db_set apache/server-name arctic.org
+ _db_cmd 'SET apache/server-name' arctic.org
+ printf '%s\n' 'SET
Package: rtorrent
Version: 0.6.0-1
this patch was extracted from upstream svn. it fixes a segfault which
occurs with libcurl 7.15.5. it tends to hit me when i have lots of
torrents going.
-dean
Index: src/core/curl_stack.cc
===
notice that tail -f works fine when the file is not marked append only...
now don't ask me why O_NONBLOCK is denied on append only files... but it
is...
I take it that there is no bug in the tail program, then.
there is absolutely a bug in the tail program... this was not a problem in
package: netbase
version: 4.33
spot the bug in /etc/init.d/networking:
process_options() {
[ -e /etc/network/options ] || return 0
log_warning_msg /etc/network/options still exists and it will be IGNORED!
Read README.Debian of netbase.
}
there should be a return 0 after the
Package: iproute
Version: 20080725-2
i did:
sudo apt-get build-dep iproute
apt-get source iproute
cd iproute-20080725-2
fakeroot ./debian/rules binary
and it fails:
...
/usr/share/texmf-texlive/dvips/base/texps.pro
/usr/share/texmf-texlive/dvips/base/special.pro
one further change to consider -- run /etc/cron.daily/ntp-server as user
ntp instead of root. maybe stick a line like this in it:
[ `/usr/bin/id -un` = ntp ] || exec /bin/su -s /bin/sh ntp $0
(or use /etc/cron.d/ntp-server which can specify a username)
-dean
--
To UNSUBSCRIBE,
Package: diff
Version: 2.8.1-10
as of 2.8.1-10 diff has reverted to braindamaged posix behaviour:
% diff -u0 a1 a2
diff: `-0' option is obsolete; use `-U 0'
diff: Try `diff --help' for more information.
zsh: exit 2 diff -u0 a1 a2
if i revert to 2.8.1-9 it works fine.
i didn't see anything
Package: libc6
Version: 2.3.2.ds1-20
i've started seeing entries like this in my syslog:
Apr 21 09:48:06 twinlark curl: gethostby*.getanswer: asked for
ethereal.net.nyud.net IN A, got type 39
type 39 is a DNAME record http://www.faqs.org/rfcs/rfc2672.html, and
DNAMEs are apparently in use by
Package: elinks
Version: 0.10.2-2
i have a cronjob which runs links -dump http://foo/; ... it worked fine
until this week when i upgraded and started getting this:
ELinks: Permission denied
i tried strace and it said this:
open(/dev/stdin, O_RDONLY|O_NOCTTY) = -1 EACCES (Permission denied)
i
Package: mdadm
Version: 2.6.2-2
it is EXEPTIONALLY DANGEROUS to replace EVERY SINGLE initrd when mdadm is
installed/upgraded.
you pretty much guarantee that any problem will produce an unbootable
system -- especially if root is on md.
as has just occured to me.
in the past in this situation
Package: libtorrent10
Version: 0.11.8-1
between 0.11.7 and 0.11.8-1 i started getting regular crashes starting
with:
** glibc detected *** /usr/bin/rtorrent: double free or corruption
(!prev): 0x0b0952b0 ***
this is on amd64.
i looked at the known issues page and it requires
damn... -fno-strict-aliasing isn't enough to fix the crash i started
seeing in 0.11.8. i built my own package, but saw a crash within 24h.
-dean
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
i rebuilt 0.11.7-1 from source (fetched from snapshot.debian.org) and it
seems not to be crashing (crashes were occuring in under a day before and
i've had 0.11.7-1 going for 2 days)... so this really is a 0.11.7-1 -
0.11.8-1 regression. i'm going to upgrade my gcc/etc to latest bleeding
edge
Package: zsh
Version: 4.3.4-23
upgrading from 4.3.4-19 to 4.3.4-23 caused zsh to be removed from
/etc/shells... i have a nightly cron job which looks for users with
invalid shells and it picked up this change last night after i did the
aforementioned upgrade yesterday.
-dean
--
To
Package: alpine
Version: 0.+dfsg-1
this is a pine 4.64 - alpine 0. regression. when a message with long
lines is piped through an external command the lines are truncated. i see
no options for scrolling the display or avoiding the truncation. note
that regular message viewing wraps
On Fri, 28 Sep 2007, martin f krafft wrote:
also sprach dean gaudet [EMAIL PROTECTED] [2007.09.28.0230 +0100]:
it is EXEPTIONALLY DANGEROUS to replace EVERY SINGLE initrd when mdadm is
installed/upgraded.
Please STOP SCREAMING and look at the existing bugs before you reply
new ones
Package: aircrack-ng
Version: 1:0.6.2-6
the aircrack-ng should Depend on the wireless-tools package... since the
airmon-ng script requires iwpriv/iwconfig.
thanks
-dean
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Package: htdig
Version: 1:3.2.0b6-3
i'd rather not get this message every day from cron:
/etc/cron.daily/htdig:
/etc/cron.daily/htdig: line 22: 1723 Terminated lockfile-touch
/var/run/htdig.cron
the patch below should quiet things.
-dean
--- etc/cron.daily/htdig.dpkg-orig
On Mon, 1 Jan 2007, Mark Brown wrote:
On Wed, Sep 06, 2006 at 05:51:39PM -0700, dean gaudet wrote:
note that this define wasn't necessary on 32-bit x86 because there's
custom 32-bit assembly which uses unaligneds even more aggressively than
the C code does even when given UNALIGNED_OK
On Mon, 1 Jan 2007, dean gaudet wrote:
and i can't even reproduce my results... here's the averages of the user
cpu seconds for 10 runs of minizip -9o a.zip linux-2.6.19.tar:
baseline -DUNALIGNED_OK
k8 revF26.62 26.59
core2 28.43 28.44
you know, gzip -9
i believe these 3 bugs are the same problem.
when the ldap server closes the connection during a response, libnss-ldap
doesn't really notice at all... it returns an error code to the caller but
doesn't pay attention to the error code itself. then nscd (or whatever
other caller is involved)
Package: watchdog
Version: 5.2.6-6
i think watchdog should be amongst the earliest to start services... who
knows if some startup script will hang. starting at 89 seems pretty late.
and... i'd really like to use nowayout=1, but if i do that then it causes
some serious cramps in my shutdown
Package: gsl
Version: 1.8-3
i did an apt-get build-dep libgsl0-dev prior to trying to build from
source... and it mostly succeeded except its looking for tex, dvips, and
ps2eps binaries.
so it seems gsl should Build-Depends: tetex-bin, ps2eps
thanks
-dean
--
To UNSUBSCRIBE, email to [EMAIL
On Sun, 11 Feb 2007, Dirk Eddelbuettel wrote:
Which distribution do you use: testing or unstable? I'd be surprised if
there had been 'unannounced' changes in these lately.
unstable
Installing the build-deps
...
- Considering tetex-extra
- Trying tetex-extra
- Considering texinfo
Package: shadow
Version: 1:4.0.18.1-7
try doing vigr/vipw and then ^Z... and fg... sometimes nastiness happens
right away, sometimes it seems to take a few ^Z/fg cycles.
grep -r for WUNTRACED you'll see the vipw.c code differs from the other
instances... patch below.
-dean
p.s. i use zsh..
On Sat, 18 Nov 2006, Michael Prokop wrote:
So can you please provide the necessary steps to reproduce the problem?
iirc it doesn't happen on every file 4GB.
try between a 32-bit and a 64-bit host -- that's when it was hitting me
the worst.
-dean
--
To UNSUBSCRIBE, email to [EMAIL
Package: nmh
Version: 1.2-1
in 1.2-1 post(8) is segfaulting (amd64)... doesn't happen with same config
on 1.1-release-4.
if i get a chance i'll grab a gdb backtrace... but maybe this strace will
help.
oh maybe my mts.conf will help too:
# grep -ve '^#.*' -e '^$' /etc/nmh/mts.conf
mts: smtp
go ahead and remove it. this package does not support apache2. (i might
be listed as maintainer but i didn't upload it and know nothing about
being a debian package maintainer.)
-dean
On Fri, 15 Jun 2007, Gerfried Fuchs wrote:
Package: libapache-mod-iptos
Severity: serious
Version: 1.1-1
Package: openssh-server
Version: 1:4.6p1-1
4.6p1 generates *lots* of log spam like so:
Jun 18 00:16:23 twinlark sshd[18923]: error: channel 0: chan_read_failed for
istate 3
Jun 18 00:16:23 twinlark sshd[18923]: error: channel 0: chan_read_failed for
istate 3
there's a fix upstream:
i've finally tracked this down:
nmh code assumes strcasecmp accepts NULL arguments.
for portability reasons sbr/strcasecmp.c defines str[n]casecmp functions
which do accept NULL arguments.
/usr/include/string.h declares strcasecmp:
extern int strcasecmp (__const char *__s1, __const char
Package: spamassassin
Version: 3.2.1-1
the following fails:
apt-get source spamassassin
cd spamassassin-3.2.1
fakeroot ./debian/rules binary
fakeroot ./debian/rules clean
because it can't deapply 10_change_config_paths ...
for one thing 10_change_config_paths includes an INSTALL.orig...
Package: linux-image-2.6.12-1-686-smp
Version: 2.6.12-7
usb audio worked in -6... but as of -7 i'm getting this in dmesg when the
module is inserted:
snd_usb_audio: Unknown symbol __compound_literal.170
snd_usb_audio: Unknown symbol __compound_literal.89
snd_usb_audio: Unknown symbol
Package: libpam-modules
Version: 0.79-2
if /etc/environment does NOT exist then logins/etc fail with a pam_setcred
critical error. (this is NOT a repeat of the other bugs related to
/etc/environment just fixed in 0.79-2 :)
-dean
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject
On Sun, 17 Jul 2005, Michael Berg wrote:
The latest libpam-umask (0.02) encounters a SIGSEGV (Segmentation fault)
this is because of some unfortunate code that i don't even think should be
in the module... it's a result of the per-user umask support. it
segfaults for any user which does
here's my patch.
sorry it's large because i think that per-user umask should be optional,
so i've done most of the work to make that happen... now you have to
specify user as an argument to pam_umask.so... as it happens it has to
be the first argument, because arguments are processed
Package: rssh
Version: 2.2.3-2
rssh logs about everything it's read from the config file... i'm getting
6+ lines every single rssh session. this is maybe useful when you're
debugging, but it's certainly too much information to log at LOG_INFO
setting every single session... this patch lowers
Package: at
Version: 3.1.9
when launching atd it's preferable to clean the environment -- in
particular things like SSH* env vars or even LOGNAME can leak into the
environment of the daemon. i've seen this cause problems such as atd
sending mail as 'dean' rather than 'root' when i restart it.
Package: clamav-freshclam
Version: 0.87-1
please add a MAILTO=root at the top of /etc/cron.d/clamav-freshclam so
that any output from freshclam failures goes to root rather than to the
clamav user...
thanks
-dean
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe.
ah, wait i see you add a clamav: root to /etc/aliases... works fine
except for those of us using MTAs which don't support /etc/aliases. i'm
satisfied enough making a local mod to my own crontab then... go ahead and
close this out, sorry.
-dean
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
Package: fail2ban
Version: 0.5.2-1
if you ask fail2ban to use SYSLOG for logging it'll try to log to
localhost:514 ... which isn't typically enabled on a debian system.
the below patch fixes this to use /dev/log, and allows the user to
optionally change the syslog-facility in the config file
Package: openssh-server
Version: 1:4.1p1-4
openssh 4.x now tries to append to /var/log/btmp (on bad passwords for
example), but it's excessively anal about the permissions on that file. it
doesn't permit group or other to have any of read/write/execute.
the default debian setup is this:
On Thu, 28 Apr 2005, Konrad Podloucky wrote:
Apparently SELinux uses extended attributes and rdiff-backup tries to
clear all attributes from temp files it creates during transfer (at least
that's what I concluded). However the security.selinux attribute can not
be removed (at least not by
the following patch fixes the chroot problem. it retains cap_sys_chroot
for a few lines longer in the code -- note there is a subsequent call
already in the code which removes all capabilities except cap_sys_time.
-dean
--- ntp-4.2.0a+stable/ntpd/ntpd.c.orig 2005-06-29 14:01:31.0
Package: ntp
Version: 1:4.2.0a+stable-8
gcc generates several type punning warnings while compiling ntp and it's
probably best to disable strict-aliasing to avoid the possibility of
incorrect optimisations. i didn't study the source to see if the type
punning could be avoided... that should
Package: ntp-server
Version: 1:4.2.0a+stable-8
i'd prefer to not have local modifications to /etc/init.d/ntp-server ... i
add -L -i /var/chroot/ntpd on my boxes. the patch below adds support
for /etc/default/ntp-server which allows the OPTIONS and RUNASUSER to be
modified.
i also made the
Package: tcpdump
Version: 3.9.0.cvs.20050614-1
this problem has been submitted upstream (tracking info at
http://sourceforge.net/tracker/index.php?func=detailaid=1232347group_id=53066atid=469575)
the -A flag no longer prints the ascii minus the hex like it did in
3.8.x... while i'm a bit
On Tue, 5 Jul 2005, Romain Francoise wrote:
dean gaudet [EMAIL PROTECTED] writes:
this problem has been submitted upstream (tracking info at
http://sourceforge.net/tracker/index.php?func=detailaid=1232347group_id=53066atid=469575)
You might want to ask on -workers, I'm not sure anyone
Package: login
Version: 1:4.0.3-36
this bug has been introduced since 1:4.0.3-35 ... perhaps related to the
fix for #314727.
with zsh as your shell, this sequence is busted:
dotlark:~% su -m
Password:
[EMAIL PROTECTED]:~# suspend
zsh: suspended su -m
i'm pretty sure this is because bash and tcsh create their own process
group at startup, and zsh doesn't... so zsh shares the same process group
as the su process.
suppose we have this pstree fragment:
zsh(4782,dean)---su(4788,root)---zsh(4789)
both pids 4788 and 4789 have pgrp 4788.
ok this is gross... but this seems to fix the problems. at first i tried
just adding the setpgrp... but with that the su'd zsh doesn't ever seem to
wake up.
so i threw in the TIOCSPGRP calls to pass the tty pgrp to the su'd zsh...
and that fixes it.
it looks like bash calls TIOCSPGRP
On Mon, 11 Jul 2005, Alexander Gattin wrote:
As I already said, I'd just prefer to block/ignore several signals
like TSTP. Most probably I'll do the same as in upstream --
block everything (except TERM and ALRM) until exit...
yeah after i stopped hacking and went to bed this popped into my
Package: fail2ban
Version: 0.8.2-3
fail2ban 0.6 supported a syslog-facility config option which controlled
the facility for syslog messages... 0.8.2-3 does not support this. i had
to edit /usr/share/fail2ban/server/server.py in order to change LOG_DAEMON
to LOG_AUTH.
-dean
--
To
Package: fail2ban
Version: 0.8.2-3
when connecting with ssh keys, no password, sshd logs:
May 18 05:08:45 twinlark sshd[5681]: Failed none for dean from 10.1.1.1 port
37262 ssh2
May 18 05:08:45 twinlark sshd[5681]: Found matching RSA key:
May 18 05:08:45 twinlark sshd[5681]: Found matching
Package: apt-listchanges
Version: 2.82
when apt-listchanges encounters an error (such as the now infamous
database /var/lib/apt/listchanges.db failed to load. error) it continues
without confirmation even if confirm=1 is in the etc file. i think
apt-listchanges should always ask for
On Fri, 24 Feb 2006, Jason Boxman wrote:
I tired it, but on my system I get a segmentation fault from mdadm when it
tried to assemble. Assembling with the usual
`mdadm --assemble /dev/md0 /dev/hde1 /dev/hdg1` works.
[EMAIL PROTECTED]:~$ uname -a
Linux faith 2.6.16-rc3-20060212 #1
On Tue, 14 Feb 2006, Guus Sliepen wrote:
tags 351185 + wontfix
thanks
On Thu, Feb 02, 2006 at 07:13:33PM -0800, dean gaudet wrote:
i think /etc/init.d/ifrename should use the -t option... which allows
devices to take over already assigned device names... i just had an
unstable box
Package: initscripts
Version: 2.86.ds1-11
i have EDITMOTD=no in my /etc/default/rcS ... yet the new
initscripts.postinst insists on turning my motd into a symlink. that
counts as an edit in my books... please leave the file alone, period.
thanks
-dean
--
To UNSUBSCRIBE, email to [EMAIL
On Tue, 14 Feb 2006, Guus Sliepen wrote:
The manual page says this about the -t option:
-t Enable name takeover support. [...] This works only with kernel
2.6.X and if the other interface is down. Consequently, this is not
compatible with Hotplug. [...] In any case, name swapping and
Package: vorbis-tools
Version: 1.1.1-3
oggenc can't handle files = 2GiB:
ERROR: Cannot open input file 20060325-1959.wav: File too large
-rw-r--r-- 1 dean music 2560491564 Mar 26 00:01 20060325-1959.wav
fix is easy... add -D_FILE_OFFSET_BITS=64 to CFLAGS in debian/rules...
i've test rebuilt
Package: yaird
Version: 0.0.12-3
it's generally preferable to tell mdadm to scan the partitions list to
find raid component devices by UUID rather than specify the device list
explicitly...
for example with an explicit device list you can end up with boot failures
if one of the devices is
Package: ifrename
Version: 27+28pre13-1
i think /etc/init.d/ifrename should use the -t option... which allows
devices to take over already assigned device names... i just had an
unstable box which was working ok with an /etc/iftab that explicitly
specified eth0/eth1 mac addresses.
then i
fyi -- linus has accepted a patch into 2.6.16 which fixes the kernel's
behaviour in this case. the kernel was requiring O_APPEND in F_SETFL on
append-only files even if the file was opened for read-only... now it
requires that the F_SETFL isn't trying to change the O_APPEND flag on an
mdrun does not respect the device minor numbers at all... this can totally
mess up the ordering of md devices and screw with static mounts (note that
while mounting by label/uuid is preferable it isn't always available --
for example XFS external log partitions cannot be specified by label or
Package: vorbis-tools
Version: 1.1.1-5
hi... thanks for trying to fix bug#359068 ... but unfortunately your fix
didn't work :) i think after seeing the patches below you might want to
consider the easy fix i suggested in the bug... changing CFLAGS directly.
there are a few problems... one is
Package: librsync1
Version: 0.9.7-1
i wonder if you would consider applying this patch from upstream to fix
problems which occur with files over 4GB.
see:
https://sourceforge.net/tracker/index.php?func=detailaid=1439412group_id=56125atid=479441
and
Package: graphviz
Version: 2.8-1
hey there... there are some dangling symlinks in the graphviz package:
/usr/lib/graphviz/lua/gv.so - libgv_lua.so
/usr/lib/graphviz/perl/gv.so - libgv_perl.so
/usr/lib/graphviz/python/_gv.so - libgv_python.so
/usr/lib/graphviz/ruby/gv.so - libgv_ruby.so
Package: initramfs-tools
Version: 0.59b
hi... if you're going to use mdrun /dev and launch all arrays during the
initrd then you should also include mdadm.conf -- because otherwise you
may end up starting arrays on the wrong minors... or even worse, if
someone is using partitioned md then
Package: coreutils
Version: 5.93-5
bug#339400 was closed prematurely...
notice that tail -f works fine when the file is not marked append only...
now don't ask me why O_NONBLOCK is denied on append only files... but it
is...
I take it that there is no bug in the tail program, then.
On Wed, 7 Jun 2006, Jonas Smedegaard wrote:
Then I'll just lean back and wait for you to do the hard work :-D
ok i've got a partial solution... see below.
i say partial because the following scenario is not quite ideal:
- md initially on /dev/sda1 /dev/sdb1 ... build initrd
- another disk
reopen 211858
tag 211858 patch
thanks
hi... i think there's been some confusion w.r.t. bug#211858. it could be
my fault, or it could just be that the package changed again between 2004
and 2006... anyhow the current state of the sysklogd package is that this
bug exists -- a regular user can
[i've cc'd the md/mdadm maintainer so he can chime in if i'm making an
error... Neil if you want to see the entire thread it's visible at
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=351183.]
On Tue, 2 May 2006, Jason Lunz wrote:
I agree. Unfortunately, telling mdadm to scan other
On Tue, 6 Jun 2006, Jonas Smedegaard wrote:
It is wrong to assume a purely no-older-than-etch system: Imagine the
process of upgrading from sarge to etch...
aha... now i understand :)
what you really need is some sort of weak dependency which implies that
if the package is already installed
On Wed, 7 Jun 2006, Jonas Smedegaard wrote:
...
Well, the above puts the test at boot time. That's ugly IMHO.
I'd prefer resolving mdadm features at build time. Something like this:
* Resolve mdadm capabilities in Plan.pm
* Set some variable (not the version, but flags each capability)
Package: cpuburn
Version: 1.4-23
there's another burn-in program which is based on gromacs inner loops
(such as those used in [EMAIL PROTECTED]) which tends to drive up cpu
temperatures even more than the programs already in cpuburn package. it
has the additional advantage of checking its
Package: bind9
Version: 9.3.2-2.1
there's several type-punning warnings while building bind9 9.3.2-2.1... i
don't know if these are bugs or not, but it's probably best to add
-fno-strict-aliasing to CFLAGS until upstream deals with them (otherwise
there could be some subtle hard to discover
Package: gzip
Version: 1.3.5-14
please define UNALIGNED_OK when building the amd64 target... unaligneds
are very inexpensive on all intel and amd cpus.
i benchmarked gzip -9 on linux-2.6.17.tar with this define and i see a
2.5% speedup on p4, a64, and a 9% speedup on core2.
note that this
1 - 100 of 119 matches
Mail list logo
