Bug#548951: ftp.linux.org.tr origin update
Hi, Simon Paillard demis ki:: > > To ensure an european mirror is returned, can you paste the result of > 'dig -ttxt +short security.debian.org' ? > It returns "AS view", I assume this is "Asia" and that explains some of the slowness, connections between Turkey and Asian countries are slow, whereas European connections are way faster. Even US connections may be faster than Asia. Can this be result of using Maxmind GeoIP database, that lists Turkey in Asia, and that represents a very inaccurate network topology. MirrorBrain has manually "fixed" it, we should fix it in debian too. BTW, MirrorBrain (mirrorbrain.org) is a great software, very efficient, and I suggest debian mirror team considers switching to it. [ Shall I open a new bug about this ? ;) ] >>> You may sync /debian-cd/ from ftp.tr.debian.org as well. >> I also changed the CD mirror origin to ftp.tr.d.o > > Thanks. > Could you please add a trace file with your name once the CD sync is > finished ? > The mirror script updates the trace file after each run. This has been the case for about a year now. > Thanks for mirroring Debian and best regards. > Thank YOU for managing the mirrors. :) -- Gokdeniz Karadag -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#548951: ftp.linux.org.tr origin update
Simon Paillard demis ki:: > severity 548951 minor > thanks > > On Tue, Sep 29, 2009 at 10:21:49PM +0300, Gokdeniz Karadag wrote: >> Package: mirrors >> >> At ftp.linux.org.tr mirror, we set up push mirroring from ftp.tr.debian.org, >> for debian and debian-volatile archives. > > About /debian-security/, since we don't advertise mirrors, you should > drop this mirror (or at least sync from ftp.tr.debian.org if you use > this security mirror for your own needs). > > You may sync /debian-cd/ from ftp.tr.debian.org as well. > I also changed the CD mirror origin to ftp.tr.d.o I do not intend to drop the security mirror, as the current DNS round robin method used by security.debian.org is suboptimal, and results in very slow package downloads. A client inside any Turkish university can download packages with several MB/s speed from our mirror, wheras a security.debian.org mirror results in <100K/s speed, as the bottleneck is international connections to/from Turkey. Thanks. -- Gokdeniz Karadag -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#548951: ftp.linux.org.tr origin update
Package: mirrors At ftp.linux.org.tr mirror, we set up push mirroring from ftp.tr.debian.org, for debian and debian-volatile archives. I am sending this update to change our origin mirror in the mirror checker listings. -- Gokdeniz Karadag ftp.linux.org.tr -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#529867: [Ftp] Re: Bug#529867: mirror listing update for ftp.linux.org.tr
Simon Paillard demis ki:: > ... > Thanks, updated. > Thank you. > Could you please add your own trace file at the end of the script that > sync ISO in http://ftp.linux.org.tr/debian-cd/project/trace/ ? > > By the way, the list considers you sync ISO from ftp.gr.d.o, is that > still true ? > > That helps a lot the investigation in case one mirror is not up to date. I set up cron to run "date -u > project/trace/ftp.linux.org.tr" after every sync. Our cd mirror source has changed to cdimage.debian.org as ftp.gr.d.o doesn't have amd64 isos. > > Best regards. > -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#506703: amavis-stats: remaining issues with newly uploaded package
Hello,
Sorry to bug again but some issues have remained for the uploaded amavis stats
package.
1. Dependencies are needed on "libunix-syslog-perl" and
"librrdp-perl". Dependency on "php4-rrdtool" must be removed as that package
does not exist anymore. Relevant lines in debian/control must be:
Depends: ${misc:Depends}, librrds-perl, librrdp-perl, libunix-syslog-perl,
rrdtool, wwwconfig-common, adduser
Recommends: amavis, php5-gd
2. amavis-stats looks for a non-existing config file "/etc/amavis-stats.conf"
at line 1120 in the sbin/amavis-stats.
New location is /etc/amavis-stats/amavis-stats.conf
relevant configure option --sysconfdir=/etc/amavis-stats
3. The package installs a cronjob but amavis-stats is a daemon now.
There must be a basic init script: /usr/sbin/amavis-stats takes start|stop as
parameters.
4. Config file must change the user to amavis-stats as data directories are
owned by that user.
$daemon_user = 'amavis-stats'; # (no default; customary: www-data)
$daemon_group = 'amavis-stats';
Note : in debian/rules file, I tried stating "--with-user=amavis-stats"
"--with-group=amavis-stats" parameters and "--disable-web" parameter to disable
checking the presence of this user at build time, still the "install" command
required the user.
5. The config file must have the following line. the original looks for a
amavis.log which does not exist
$scan_logfile = "/var/log/mail.info"
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#506703: deps for backend
Here are two missing dependencies for the "backend" part of the package libunix-syslog-perl librrdp-perl (in addition to librrds-perl) Also, I have not stated it clearly in previous e-mail, the config file must be patched to read as following, amavis debian pkg outputs are in there. $scan_logfile = "/var/log/mail.info" -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#506703: new version (0.1.22) for amavis-stats
Hi,
I have just had the time to test the package. Thanks for providing it. My
report will be rather long, sorry for it. But at the end the thing works now :D
If I remember correctly, amavis-stats prefixes itself to /usr while building.
And It seems that the package is built with --prefix=/usr.
In line 1120 of /usr/sbin/amavis-stats it looks for config file at
'/etc/amavis-stats/amavis-stats.conf'. Also, it installs
/usr/var/cache/amavis-stats and /usr/var/lib/amavis-stats directories.
There must be a configure option for the prefix.
Also it puts files in "/usr/doc", no other package does this according to
debian policy. and /usr/doc/amavis-stats.txt is in fact a man file -with
horrible formatting :(
This program really sucks but there is no ready-made alternative.
The data directories (/var/{lib,cache}/amavis-stats) must be www-data owned.
And I cannot remember what original permissions were on /var/log/mail.info, but
amavis-stats must be able to read it.
The new version defaults in a daemon-like operation. So the cron.d file is
unnecessary, instead an initscript must be installed, that will directly call
/usr/sbin/amavis-stats with start|stop etc. Be aware that process renames
itself as `Amavis-Stats`
/usr/share/amavis-stats/amavis-stats.alias.conf is similar to the apache.conf,
with added expiration instructions, it should not be present in installed
package.
In my setup, /etc/amavis-stats/apache.conf was not linked under apache's conf.d
Also I had to fix /usr/share/amavis-stats/img link, as the current package has
an extra "/usr/" prefix in all directories, in the corrected version, that link
should be pointing to correct place.
Some dependencies are missing; a dep on php and php-gd must be present for php
files. I am not very comfortable with this. People do not install web servers
on their mail servers. I would not prefer such a dep, instead I copy rrd files
over to a apache installed machine. Can you provide a backend package that
produces rrd files, and a frontend that produces graphics ?
After all these adjustments, I managed to get graphics of virus, spam and other
activity. Yay!
Actually, the "business1" theme seems like professional thingy :)
Thanks for your efforts.
Barry deFreese demis ki::
> Barry deFreese wrote:
>> Gokdeniz Karadag wrote:
>>> Hi,
>>>
>>> I can test the package. Where can I find it ?
>>>
>>> - Gokdeniz Karadag
>>>
>>> Barry deFreese demis ki::
>>>
>>>> Hi,
>>>>
>>>> I have a package of 0.1.22 ready to go but I am not overly familiar
>>>> with
>>>> using it. Is there any chance you could do some testing for me?
>>>>
>>>> Thanks!
>>>>
>>>> Barry deFreese
>>>> Debian QA
>>>>
>>>
>>>
>> Gokdeniz,
>>
>> Thanks. I have put the .deb here for now:
>> http://people.debian.org/~bdefreese/amavis-stats/amavis-stats_0.1.22-1_all.deb
>>
>>
>> Unfortunately it isn't apt-gettable so you will have to wget and use
>> dpkg -i. If you don't know how to do that, please let me know. I'm
>> sure there is some extraneous stuff in the package since the way it is
>> configured seems to have changed significantly. In fact the
>> apache.conf file that is shipped is probably no good anymore as well
>> as the cronjob probably needs modification. Any help/suggestions you
>> could give there would be greatly appreciated.
>>
>> Thanks,
>>
>> Barry
>>
> Gokdeniz,
>
> Hi, sorry to bug you, just curious if you've had a chance to test this
> version out?
>
> Thanks!
>
> Barry deFreese
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#498716: [Ftp] Re: Bug#498716: Bug#498716: mirror listing update for ftp.linux.org.tr
Simon Paillard demis ki:: > Just another notice / reminder : > * HTTP parameters to list the size and last modified > * ftpsync script to be used, see http://debian.org/mirror/ftpmirror#how > Both are done now. Sorry for the rather long delay. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#526835: The same error on lenny/squeeze system
I get the same error on an lenny system, which pulls cherokee packages from squeeze via apt-preferences. The upgrade waas from 0.99.9-1 to 0.99.11-1. After libcherokee-config0 was upgraded to 0.99.11-1, I got the error. When I tried to remove it, the same error was also produced. I'm pasting it below: Removing libcherokee-config0 ... Usage: update-python-modules [-v] [-c] package_directory [...] update-python-modules [-v] [-c] package.dirs [...] update-python-modules [-v] [-a|-f|-p] update-python-modules: error: /usr/share/python-support/libcherokee-config0.private is not a directory dpkg: error processing libcherokee-config0 (--remove): subprocess pre-removal script returned error exit status 2 Usage: update-python-modules [-v] [-c] package_directory [...] update-python-modules [-v] [-c] package.dirs [...] update-python-modules [-v] [-a|-f|-p] update-python-modules: error: /usr/share/python-support/libcherokee-config0.private is not a directory dpkg: error while cleaning up: subprocess post-installation script returned error exit status 2 Errors were encountered while processing: libcherokee-config0 E: Sub-process /usr/bin/dpkg returned an error code (1) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#506703: new version (0.1.22) for amavis-stats
Hi, I can test the package. Where can I find it ? - Gokdeniz Karadag Barry deFreese demis ki:: > Hi, > > I have a package of 0.1.22 ready to go but I am not overly familiar with > using it. Is there any chance you could do some testing for me? > > Thanks! > > Barry deFreese > Debian QA -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#502798: wrong information
My previous scenario was bogus because the working virtual machine had the etch kernel in its config file. when I changed it to use lenny's xen kernel, it showed the symptomps. sorry for the misinformation and bugspam. -- Gokdeniz Karadag -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#502798: upgrade from etch
To add some extra information; I was using etch domU's, created by debian's xen-tools. I have upgraded two virtual machines to lenny. To tell the situation let's call the dom0's hostA, hostB, hostC and domU's as virt1 and virt2. virt1 was copy of virt2 with some additional server packages. virt1 was on hostA, they were upgraded to lenny at the same time, then rebooted. It all runs fine; - virt1 doesn't have a getty listening to hvc0 but console works fine. - virt1 doesn't have udev installed but no problems while logging in with SSH(it uses ttyp0 as TTY) virt2 was on hostB. virt1 was upgraded but hostB was not. hostC was upgraded to lenny in isolation, no virtuals on this one. Then virt2 was moved to hostC. Altough system packages of virt1 and virt2 were all the same, - virt2 had the SSH connectivity issue because of a not mounted devpts fs on /dev/pts - virt2 had the console issue because of absence of a getty on hvc0 I fixed virt2 by installing udev and creating a getty for hvc0 in /etc/inittab "0:2345:respawn:/sbin/getty 38400 hvc0" But I wonder why virt1 continued working and I'm willing to perform tests, do experiments, share relevant configs etc. to shed some more light on the issue. -- Gokdeniz Karadag -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#521190: aptitude state-bundle command man pages should be informative
package: aptitude Hi, Man pages for aptitude-create-state-bundle and aptitude-run-state-bundle do not explain their purpuse clearly. Man pages state "what" they do ( "unpacks bundle, invokes program" etc. ) but does not hint "why" you should be doing it. It should detail what the user will achive by running this pair of commands and is it a complete replacement of "dpkg --set-selections < file && apt-get dselect-upgrade" -- Gokdeniz Karadag -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#520122: udev configuration fails at start-stop-daemon
package: udev version: 0.125-7 While upgrading from etch to lenny udev package configuration exited with this error message "subprocess post-installation script returned error exit status 2" I edited /var/lib/dpkg/info/udev.postinst file by hand to pinpoint the issue, while running "dpkg --configure udev". The problem is, the following command in restart_udevd exits with value 2. start-stop-daemon --stop --name udevd --oknodo --quiet --retry 5 The message is "Program udevd, 1 process(es), refused to die." When I see ps output, there is a defunct udev. root 1398 1 0 2008 ?00:00:00 [udevd] When I run udev manually by "udevd --debug", the process runs fine. Don't know about the functionality though, the server is far away to test by plugging some usb devices. When I tried to continue "dpkg --configure udev" The second udev process started manually by me became defunct. The debug output ends like .. [28758] parse_file: reading '/etc/udev/rules.d/z60_xen-backend.rules' as rules file [28758] main: initialize max_childs to 256 [28758] main: initialize max_childs_running to 72 gazali:~# I think postinst script can just pass the error and continue. -- Gokdeniz Karadag -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#516823: try memtest
Hi Csányi, I suspect you have a corrupt memory or some other hardware issue. First check if your disk is not full by issuing "df" command. Then select memtest from grub menu and leave it for a few hours of testing, you will probably see errors. -- Gokdeniz Karadag -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#409091: Include next release key in keyring package
Hi, Adeodato Adeodato Simó demis ki:: > Yes, we plan to do this, will be included in the next point release of > Etch (and a key for Squeeze will be included in the next point release > of Lenny as well). Thanks for informing. Well, this thing is obvious and probably something prevented it from happening at lenny release, but may I suggest a policy change about including the squeeze key in lenny before squeeze is released ? [and of course similar for next releases] -- Gokdeniz Karadag -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#409091: Include next release key in keyring package
This bug is for sarge->etch upgrade, but the problem is still present in etch->lenny upgrade. The debian-archive-keyring package in etch-volatile contains the key for lenny. But the package in etch main does not contain key for lenny. Volatile repository was not enabled by default in etch, so for a smooth upgrade, lenny key must be added to debian-archive-keyring package in etch main repository. -- Gokdeniz Karadag -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#510128: make-memtest86+-boot-floppy modified to use variables
Package: memtest86+
Here is a modified make-memtest86+-boot-floppy to use variables instead of
hardcoded paths. Also incorporated patch in bug #451050. This one can be more
easily modified to fit into other distributions.
#!/bin/sh
#
# Script for making a memtest86 boot floppy using GRUB as bootloader
#
# (c) 2003 Peter Loje Hansen
# - original version
# (c) 2004 Yann Dirson
# - added parameters
# - ability to work on a floppy image instead of a real floppy
# - adapted patches from Martin Koeppe , to use
#mtools and install full grub
# TODO:
# - add a flag to generate a default boot entry for (hd0)
set -e
MEMTEST=/boot/memtest86+.bin
FLOPPYIMAGE=/dev/fd0
GRUBBIN=/usr/sbin/grub
GRUBLIB=/usr/lib/grub
MFORMAT=/usr/bin/mformat
arch=$(dpkg --print-architecture)
case "$arch" in
i386|?*-i386) GRUBARCH=i386;;
amd64|?*-amd64) GRUBARCH=x86_64;;
*) error "Unsupported architecture: $arch";;
esac
error()
{
echo >&2 "$0: $*"
exit 1
}
needsarg()
{
[ $1 -ge 2 ] || error "syntax error"
}
[ -d $GRUBLIB ] || error "Can't find $GRUBLIB - did you install a recent grub
package (0.95+cvs20040624 or later) ?"
[ -x $MFORMAT ] || error "Can't find mformat - did you install the mtools
package ?"
while [ $# -gt 0 ]
do
case "$1" in
--help) echo "$0 [--memtest $MEMTEST] [--floppyimage $FLOPPYIMAGE]"; exit 0
;;
--memtest) needsarg $#; MEMTEST="$2"; shift ;;
--floppyimage) needsarg $#; FLOPPYIMAGE="$2"; shift ;;
*) error "syntax error" ;;
esac
shift
done
MOUNTPOINT=$(mktemp -d)
if [ -b "$FLOPPYIMAGE" ]
then
FINALDEV="$FLOPPYIMAGE"
FLOPPYIMAGE="$(mktemp)"
else
FINALDEV=""
fi
echo "* Creating msdos file system"
echo
if [ ! -s "$FLOPPYIMAGE" ]; then
# unless a non-empty image exists, create a blank one first
dd bs=1024 count=1440 if=/dev/zero of="$FLOPPYIMAGE"
fi
# FIXME: "-f 1440" should probably be dropped
mformat -i $FLOPPYIMAGE -f 1440 ::
mmd -i $FLOPPYIMAGE ::/boot
mmd -i $FLOPPYIMAGE ::/boot/grub
echo
echo "* Installing GRUB files"
mcopy -v -i "$FLOPPYIMAGE" - ::/boot/grub/menu.lst <
Bug#498716: [Ftp] Bug#498716: mirror listing update for ftp.linux.org.tr
Hi, Sorry for the late reply, somehow I have forgotten to send the reply I prepared then, and didn't look back. Simon Paillard demis ki:: We have released a new script set much better for sync mirroring, we advise you to use it (remember to subscribe to debian-mirrors-announce): http://lists.debian.org/debian-mirrors-announce/2008/11/msg1.html Could you apply the recommended Apache settings below : http://www.debian.org/mirror/ftpmirror ? (It's a need for the CD mirror in order to check the size and last modified date) Still applicable. We are in the middle of moving into a chroot in our host's server (Turkish National Academic Network), and the move is proceeding slowly. We do not want to bother their team with our issues, when the move is complete, we will have root access to chroot and resolve these two issues with debian mirroring. Archive-upstream: ftp.uk.debian.org CDImage-upstream: ftp.gr.debian.org Is something wrong with ftp.tr.d.o we should be aware ? Same question, if there is a problem seen from your country, tell us. There's not a problem with tr.d.o, we just wanted to have a different path to download from. Reason is to provide redundancy(somewhat) and to have a different copy if errors or file corruption occured on tr.d.o Thanks ! -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#504347: backport the fix to etch?
Hi, The second hunk of the patch[1] seems trivial enough to add to etch version of dovecot, is this possible ? [1]: http://hg.dovecot.org/dovecot-sieve-1.0/rev/8b6df1256798 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#506703: new version (0.1.22) for amavis-stats
Package: amavis-stats There is a new version of amavis-stats on the following web page; http://osx.topicdesk.com/content/view/42/59/ the package seems to be improved greatly over the version in debian repos. To begin with, it has configuration file support :) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#365530: progress
Hi, What is the status of amd64 port now ? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#503212: dovecot configuration file mislead
Package: dovecot
Version: 1.0.rc15-2etch4
This is a request to fix a bogus warning in dovecot's config file.
While trying to use dovecot as postfix sasl auth mechanism, I came across this
comment in /etc/dovecot/dovecot.conf
auth default {
.
# /etc/passwd or similar, using getpwnam()
# In many systems nowadays this uses Name Service Switch, which is
# configured in /etc/nsswitch.conf. WARNING: nss_ldap is known to be broken
# with Dovecot. Don't use it, or users might log in as each others!
# http://wiki.dovecot.org/AuthDatabase/Passwd
userdb passwd {
}
I did some research starting on the given wiki page. That page gives a
workaround for dovecot 1.0.rc23 and later, but etch has 1.0.rc15.
Following the bug referenced, I found that root of the problem was fixed in
libnss-ldap (251-7.5etch1)
# Fix race condition, which could lead to a DoS, when applications
# use pthread and fork after a call to nss_ldap Fixes: CVE-2007-5794
The link for CVE http://secunia.com/advisories/cve_reference/CVE-2007-5794/
So, there seems to be no risk in using libnss-ldap and dovecot+PAM in etch.
The comment should be removed or expanded to explain that a problem 'was'
present but fixed with libnss-ldap version 251-7.5etch1.
--
Gokdeniz Karadag
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#501313: [php-maint] Bug#501313: Bug#501313: Fixed in lenny?
Raphael Geissert demis ki:: > [Please do not send me a copy of the message, IOW: no To nor B/CC] > > 2008/10/16 Gokdeniz Karadag <[EMAIL PROTECTED]>: > [...] >> There is a problem in etch. >> >> In etch, I have changed the gc_probability in php.ini, >> I gave a high probability/divisor ratio (about 1/10, also tried 1/1) >> and restarted apache. Garbage collection never happened. > > And what sessions path are you using? are you sure the process does > actually have enough rights to delete the files? are you sure the max > life time was reached? > The sessions are stored in a database, and there is write access to database. The webapp(Roundcube) did not have much users, there were 3000+ sessions in the database. When garbage collection was moved to the webapp code my first login initiated garbage collection and number fell to 16. >> As soon as I copied those settings into the webapp's code with iniset, >> garbage collection worked. >> >> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=267720 >> This states that gc_probability was compiled in as 0, >> """ >> * Comment out session.gc_probability in the default php.ini, as we've >> now compiled in a default of 0, allowing the cronjob to do the >> garbage collection for us instead. (closes: #267720) >> """ >> >> Changing it back may allow the setting in php.ini work. >> > > The only change that was done here was to default to 0 in php itself > (i.e. if gc_probability is not defined in php.ini it will default to > 0/off). I also guess so, but somehow the setting in php.ini does not get honored. (I restarted apache, made sure no stale process exists.) This bug was solved in roundcube by adding the ini_set call to the code, but other people may have problems with other software. - Gokdeniz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#501313: [php-maint] Bug#501313: Fixed in lenny?
Raphael Geissert demis ki:: > Hi, > > 2008/10/15 Samat K Jain <[EMAIL PROTECTED]>: >>> I cannot test with lenny as I have no installed lenny system, but if >>> gc_probability setting in php.ini is honored, then the bug is fixed. >> It is honored on lenny, yes. Though, it remains disabled (gc_probability = >> 0). >> >>> Is it possible to make php in etch aware of the gc_probability setting By >>> using the same fix? >> Someone else will have to answer that... But since lenny will become the new >> stable soon (hopefully), it may be better use of effort to prepare a switch >> to lenny than debugging this in etch. >> > > IIRC gc_probability should work in etch just like it does in lenny. It > was just turned off in the default php.ini, nothing else. > > Please confirm soon, or else I'll just close this bug as there is no real bug. There is a problem in etch. In etch, I have changed the gc_probability in php.ini, I gave a high probability/divisor ratio (about 1/10, also tried 1/1) and restarted apache. Garbage collection never happened. As soon as I copied those settings into the webapp's code with iniset, garbage collection worked. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=267720 This states that gc_probability was compiled in as 0, """ * Comment out session.gc_probability in the default php.ini, as we've now compiled in a default of 0, allowing the cronjob to do the garbage collection for us instead. (closes: #267720) """ Changing it back may allow the setting in php.ini work. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#501313: Fixed in lenny?
Hi I think you meant also this bug http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=321460 I cannot test with lenny as I have no installed lenny system, but if gc_probability setting in php.ini is honored, then the bug is fixed. Is it possible to make php in etch aware of the gc_probability setting By using the same fix? Samat K Jain demis ki:: > Do you still see this behavior in lenny? In lenny, setting > session.gc_probablity to a non-zero enables garbage collection again. > > This has been "fixed"; there's now a mention in php.ini to turn this setting > on, see bugs: > > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=388808 > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=388808 > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#501313: php5: lack of session.gc_probability breaks roundcube
Package: php5
Severity: normal
RoundCube webmail and probably some other web applications install their
own session handlers via php functions. They store sessions in their own
databases and handle garbage collection on their own.
In default etch installation, php5 ships with session.gc_probability
setting turned off. It clears "file based" sessions wtih the cronjob,
but aforementioned webapps get no chance to run their garbage
collectors. Is it possible to have both gc_probability set, and the
cronjob running in a default install ?
Also, I have experienced that setting gc_probability in
/etc/php5/apache2/php.ini had no effect ( observed with a low, divisor
and over a long time ) but setting it with the following function call
managed to run the custom garbage collection, If this is not intentionally
disabled in the package compilation, that might be a bug.
ini_set('session.gc_probability', 1);
-- System Information:
Debian Release: 4.0
APT prefers stable
APT policy: (500, 'stable')
Architecture: amd64 (x86_64)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.18-6-xen-amd64
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#383760: bash: XON/XOF flow control makes incremental search unusable
Is there any progress on this one ? I believe this bug should be fixed and there should be a backport to etch. A new user can accidentally press "ctrl+s" and cannot understand why the terminal froze. I am not a new user, I have 4+ years of sysadminning and I have found out this very recenly ( after starting to use screen very frequently, which makes pressing ctrl+s very easy because screen uses ctrl+a for every command, and 'a' is next to 's') The ubuntu bug about this topic; https://bugs.launchpad.net/debian/+source/bash/+bug/80635 -- Gokdeniz Karadag -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#480896: nscd crashes with capital letters and ldap
Package: nscd Version: 2.3.6.ds1-13etch5 we are using nscd and ldap authentication with libnss-ldap and libpam-ldap. Ldap server is sun directory server. All of our users have lowercase letters, when we try to getent or finger a username with capital letters, nscd crashes after a while, when reloading that user in the password cache. It is a strange bug, I am not really sure if this is really a nscd bug or some other part of the system, but we have consistently seen the same case in our tests. And the crashing part is nscd. During our tests we tried one uppercase letter in username, half of the name in uppercase, all uppercase letters; they all crash while reloading in password cache. A sample log can be found below. 26890: Reloading "5355" in password cache! 26890: Reloading "43466" in password cache! 26890: Reloading "101" in password cache! 26890: Reloading "6418" in password cache! 26890: remove GETPWBYUID entry "43466" 26890: remove GETPWBYUID entry "100" 26890: remove GETPWBYNAME entry "e111" 26890: remove GETPWBYNAME entry "sshd" 26890: Reloading "SELMA" in password cache! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#480835: arpwatch reports bogons even with -N option
Package: arpwatch Version: 2.1a13-2 Arpwatch package comes with "-N" as a default option, which prevents reporting of bogons. The ps output verifies that option is passwd to arpwatch, but arpwatch still reports ip addresses outside the defined network's range. I examined the code and found out that, it does not send "bogon" warning but it still sends "new station" alerts for bogons. I think proper solution would be to return from process_fddi and process_ether if the packet is bogon, without calling ent_add. Then isbogon should return 1 regardless of value of nobogon. I can prepare a patch implementing this if it will get accepted. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#465987: inetutils-talkd does not work with openbsd-inetd
Package: inetutils Version: 2:1.5.dfsg.1-2 etch installs openbsd-inetd as default inetd, but inetutils-talkd did not work with that, I have tried a few combinations in /etc/inetd.conf. The talkd package worked fine, out of the box, it sets up the inetd.conf lines. Perhaps inetutils-talkd should depend on inetutils-inetd? Altough the package names _suggests_ some connection, it should be marked as such in dependencies. I have not tested but the incompatibility with openbsd-inetd may be present in other inetutils-* packages. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#453868: Race condition in nss_ldap
Package: libnss-ldap Tags: security It is reported that there is a race condition in libnss-ldap package and it caused problems with dovecot imap/pop server. Reference [1] states that nss_ldap version 258 solves the issue. [1] http://bugs.gentoo.org/show_bug.cgi?id=198390 [2] http://wiki.rpath.com/wiki/Advisories:rPSA-2007-0255 [3] http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2007-5794 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#416120: Mistakes
Sorry, I have made mistakes and the new decision is reflected in debian. But the debian tzdata differs from the upstream for changes after 1985 in Europe/Istanbul zone. It might cause problems in past time calculations and it should be synced with the source. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#416120: Time of DST change is different for Europe/Istanbul
Package: tzdata Version: 2006p-1 Upstream 2007d version has the correct DST info for Europe/Istanbul (Asia/Istanbul) time zone. Basically, a sync with EU practices. The comment block from ftp://elsie.nci.nih.gov/pub/tzdata2007d.tar.gz states that change and I can confirm this as a Turkish citizen. After the comment I pasted the relevant part of the "europe" file from the same file. # From Steffen Thorsen (2007-03-09): # Starting 2007 though, it seems that they are adopting EU's 1:00 UTC # start/end time, according to the following page (2007-03-07): # http://www.ntvmsnbc.com/news/402029.asp # The official document is located here - it is in Turkish...: # http://rega.basbakanlik.gov.tr/eskiler/2007/03/20070307-7.htm # I was able to locate the following seemingly official document # (on a non-government server though) describing dates between 2002 and 2006: # http://www.alomaliye.com/bkk_2002_3769.htm # Rule NAMEFROMTO TYPEIN ON AT SAVELETTER/S RuleTurkey 1916only- May 1 0:001:00S RuleTurkey 1916only- Oct 1 0:000 - RuleTurkey 1920only- Mar 28 0:001:00S RuleTurkey 1920only- Oct 25 0:000 - RuleTurkey 1921only- Apr 3 0:001:00S RuleTurkey 1921only- Oct 3 0:000 - RuleTurkey 1922only- Mar 26 0:001:00S RuleTurkey 1922only- Oct 8 0:000 - # Whitman gives 1923 Apr 28 - Sep 16 and no DST in 1924-1925; # go with Shanks & Pottenger. RuleTurkey 1924only- May 13 0:001:00S RuleTurkey 19241925- Oct 1 0:000 - RuleTurkey 1925only- May 1 0:001:00S RuleTurkey 1940only- Jun 30 0:001:00S RuleTurkey 1940only- Oct 5 0:000 - RuleTurkey 1940only- Dec 1 0:001:00S RuleTurkey 1941only- Sep 21 0:000 - RuleTurkey 1942only- Apr 1 0:001:00S # Whitman omits the next two transition and gives 1945 Oct 1; # go with Shanks & Pottenger. RuleTurkey 1942only- Nov 1 0:000 - RuleTurkey 1945only- Apr 2 0:001:00S RuleTurkey 1945only- Oct 8 0:000 - RuleTurkey 1946only- Jun 1 0:001:00S RuleTurkey 1946only- Oct 1 0:000 - RuleTurkey 19471948- Apr Sun>=16 0:001:00S RuleTurkey 19471950- Oct Sun>=2 0:000 - RuleTurkey 1949only- Apr 10 0:001:00S RuleTurkey 1950only- Apr 19 0:001:00S RuleTurkey 1951only- Apr 22 0:001:00S RuleTurkey 1951only- Oct 8 0:000 - RuleTurkey 1962only- Jul 15 0:001:00S RuleTurkey 1962only- Oct 8 0:000 - RuleTurkey 1964only- May 15 0:001:00S RuleTurkey 1964only- Oct 1 0:000 - RuleTurkey 19701972- May Sun>=2 0:001:00S RuleTurkey 19701972- Oct Sun>=2 0:000 - RuleTurkey 1973only- Jun 3 1:001:00S RuleTurkey 1973only- Nov 4 3:000 - RuleTurkey 1974only- Mar 31 2:001:00S RuleTurkey 1974only- Nov 3 5:000 - RuleTurkey 1975only- Mar 30 0:001:00S RuleTurkey 19751976- Oct lastSun 0:000 - RuleTurkey 1976only- Jun 1 0:001:00S RuleTurkey 19771978- Apr Sun>=1 0:001:00S RuleTurkey 1977only- Oct 16 0:000 - RuleTurkey 19791980- Apr Sun>=1 3:001:00S RuleTurkey 19791982- Oct Mon>=11 0:000 - RuleTurkey 19811982- Mar lastSun 3:001:00S RuleTurkey 1983only- Jul 31 0:001:00S RuleTurkey 1983only- Oct 2 0:000 - RuleTurkey 1985only- Apr 20 0:001:00S RuleTurkey 1985only- Sep 28 0:000 - RuleTurkey 19861990- Mar lastSun 2:00s 1:00S RuleTurkey 19861990- Sep lastSun 2:00s 0 - RuleTurkey 19912006-
