Re: Wheezy and Sun-Java
Dear linuxers, According to this blog ( http://sylvestre.ledru.info/blog/sylvestre/2011/08/26/sun_java6_packages_re moved_from_debian_u ) from Sylvestre, responsible for maintaining sun-java*, Java did not renew it's license, and so, sun-java6-jre and sun-java-plugin will not be available on Wheezy. The author suggests: apt-get --purge remove sun-java6-jre apt-get install openjdk-7-jre in another (more recent) post. Several folks have recommended the icedtea plug-in, that's a good one. What I have done on my (squeeze) systems is, get the Sun Java tgz, and unpack it into /usr/local/stow/jre1.7.0_05 (or whatever, the tgz will create the directory), and then push it into /usr/local with stow. The browser plug-in will then be in /usr/local/lib/amd64/libnpjp2.so, you can sym-link to it from /usr/lib/mozilla/plugins, and it will then be preserved across versions if you un-stow an old one and re-stow a new one, and almost all browsers will find it there. (Konqueror won't, but Konqueror's java set-up is different anyways). I did this because I had a few things that absolutely insisted on havnig Sun Java -- it's a bit fiddly to set up (and you may have to learn about stow, but this is time well spent...), but it's reasonably robust. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201208142205.17815.rei...@bellatlantic.net
Re: java plugin, is does it have any security vulnerabilities ?
Hi All, i am running Debian stable (squeeze) and have java installed 6.26-osqueeze1 as a package. Firefox uses this package but when I asked it to check if plugins are up to date it says nope, need version 7... So am I right in thinking that v6.26 is old (and thus the advice to upgrade to 7) but does not have any known security vulnerabilities because it is part of stable ? Are you using the Debian-provided sun-java6-plugin package? My understanding is that this package is no longer updated by Debian, because of licensing restrictions introduced by Oracle affecting redistribution. There is no Debian-packaged Sun Java VM newer than 6u26. You can switch to OpenJDK, or remove your sun-java6 packages and install the JRE from the Oracle-provided binaries. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201207141823.58980.rei...@bellatlantic.net
Re: Unreversable dimmiing of display
I have Debian Squeeze installed on an ASUS EeePC 901, using the i915 video drivers and linux-image-3.2.0-0.bpo.2-686-pae from squeeze- backports. If I dim the display using the KDE power gadget a lot, or allow it to dim automatically due to low battery, it dim suddenly to a barely perceivable level and I can not restore the display to full brightness without rebooting except for a brief flicker when the display is switched off. I had an older ThinkPad that used to do this, and I never solved it, but I did find a work-around -- if I did ctrl-alt-F1 to switch to a text console, the keyboard brightness controls would work, and then when I did ctrl-alt-F7 to switch back to the X-window system, the brightness would be preserved. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201207071644.11254.rei...@bellatlantic.net
Re: UDEV identical devices. How to?
If you're set on using udev, it should be possible to find the
devices under /sys somewhere (/sys/bus/pci/...), and query them
for udev-sensitive attributes using udevinfo (or the appropriate
udevadm commands), until you find something that differs between
the two. Udev certainly can create device nodes, but I am not
an expert on that part.
But, you may find it simpler to just create static devices. It used
to be possible to create static devices in /lib/udev/devices -- at
boot-time, the device tree at that location gets copied over to
/dev, and after that, udev runs and does all the magic to create
the dynamic devices. If you know the right major and minor numbers,
and if they're consistent across reboots, and if there are no
conflicting udev devices, then the right mkdev should work.
It won't appear in /dev until reboot, but the /lib/udev/devices
entries will persist across reboots.
This is still true in the 2.6.32 kernels (squeeze), but
may not be true in the 3.0 series.
Good day to all
I am not sure which list this quistions should be send to so some
overkill maybe.
We have Debian 2.6.32 and DVB cards with two/four tunners on the each
card. We want to assign specific number in /dev/dvb/ tree for each tuner.
The problems is that from udev point of view those devices(tuners) are
absolutely equal(udefinfo gives absolutely identical output), because
tuners are placed in one dvb card.
We'v googled and found that theoretically we can assign numbers in
specific order with using ENV options in udev rules.
Here is a working example which creates symlinks:
# Create a symlinks for both tuners of Kworld device
SUBSYSTEM==dvb, ATTRS{idVendor}==1b80, ATTRS{idProduct}==e399,
ENV{kworld}!=two, ENV{kworld}=two, PROGRAM=/bin/sh -c 'K=%k;
K=$${K#dvb}; printf dvb/adapter_kw1/%%s $${K#*.}', SYMLINK+=%c
SUBSYSTEM==dvb, ATTRS{idVendor}==1b80, ATTRS{idProduct}==e399,
ENV{kworld}==two, ENV{kworld}=one, PROGRAM=/bin/sh -c 'K=%k;
K=$${K#dvb}; printf dvb/adapter_kw2/%%s $${K#*.}', SYMLINK+=%c
But we do want devices instead of symlinks, like e.g. /dev/dvb/adapterX
We'v tried such rules, but has no luck:
SUBSYSTEM==dvb, KERNELS==:04:00.0, ENV{kworld}!=two,
ENV{kworld}=one, PROGRAM=/bin/sh -c 'K=%k; K=$${K#dvb}; printf
dvb/adapter1/%%s $${K#*.}', NAME=%c, GROUP=video
SUBSYSTEM==dvb, KERNELS==:04:00.0, ENV{kworld}==two,
ENV{kworld}=one, PROGRAM=/bin/sh -c 'K=%k; K=$${K#dvb}; printf
dvb/adapter2/%%s $${K#*.}', NAME=%c, GROUP=video
Can someone give an advice how to properly construct rules to assign
devices. It woul be great with an example string :)
Mb we missing some global udev option to enable such constructions?
Tnx in advance.
--
Andrew Reid / rei...@bellatlantic.net
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/201207031829.19974.rei...@bellatlantic.net
Re: [1/2 OT] How to find the desktop near me?
Hi, The desktop installed the Sid (seems not so relevant here.) Now the monitor/screen not work; other things are totally fine. I used to login desktop first and then ssh from laptop, Now since not be able to log in, how can I find its ip ? so I can connect from laptop. or shall I just find other monitor and connect? If it's on the same subnet as your laptop, you may be able to find it via the laptop's ARP cache -- this is the mechanism the kernel uses to map controller MAC addresses to IP addresses. Alternatively, if it's gotten a DHCP assignment from your home router, you can probably go in to the router's set-up interface, and find the table of DHCP assignments it has handed out, and see if there's one you don't recognize. If all that fails, if it's a small address space (like 192.168.1.*), it s a simple matter to just ping every address in the space and see if there's a live one you didn't know about. If the machine doesn't answer ping requests, a more complicated way would be to run a port-scanner from the laptop, and see what it finds. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201207011101.34845.rei...@bellatlantic.net
Re: Reports of (Debian?) Linux kernel 2.6.32 livelocking when notified of leap second
Hello, Has anyone been seeing this sort of thing in the last 22-ish hours? http://serverfault.com/questions/403732/anyone-else-experiencing-high-rates -of-linux-server-crashes-today I saw the buzz about this on Slashdot, but as far as I can tell none of my systems are having any problems. I have several Debian squeeze 2.6.32 64-bit systems, a mix of file servers, cluster nodes, and desktop workstation systems. I'm off-site, it being the weekend, but there's no evidence of trouble. I did see a lot of log traffic at 23:59:60 UTC, so the kernels have the info. If they're going to lock up at local midnight, well, I guess I'll find out in half an hour. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201206302329.55719.rei...@bellatlantic.net
[SOLVED!!!] Weird Wifi problem -- WPA-EAP TTLS fails
There's a bit of acronym-soup here, for context check out the list archives from May 2011. I had this weird WPA-EAP TTLS problem at that time, and have now finally gotten it resolved. The key was to finally run a packet sniffer on the wireless port and watch the authentication as it went by. It turns out that the access point is asking for PEAP, not TTLS, so the EAP method-negotiation process was just failing, since the instructions I had were very clear about setting up the EAP to be TTLS-only. It's now clear that the instructions were simply wrong. The reason for mentioning this on-list is that the symptoms of client misconfiguration included, in the verbose wpa_supplicant output, a bit about ...disconnected by local choice (reason=3)..., and googling that message gets you a lot of hits, which turned out to be completely irrelevant to my issue, and I spent a lot of time on irrelevant details. Lesson 1: The local choice message can arise from a trivially misconfigured client, it's not necessarily a symptom of all the scary stuff that's out there. Lesson 2: Wireshark is awesome. I should have used it much earlier. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120011.52257.rei...@bellatlantic.net
Re: Nepomuk
I didn't ask for it and I don't want it. So why do we have to have it? It hogs the CPU and has an enormous database. I certainly do not have enough file to justify a 60Mb database. Is it possible to stop it from working in KDE? And what about nepomuk? I'd like to stop that too. Do I have to exit from KDE and use another desktop? You can kill nepomuk. To do this for one account, go to: K - System Settings - Advanced - Service Manager, and in the list of start-up services, uncheck Nepomuk Search Module. I also uncheck Update Notifier, Free Space Notifier, and Network Status Daemon. Then, in K - System Settings - Advanced - Desktop Search, turn everything off in the Basic Settings tab. You may have to log out and log back in to have it all take effect, and of course this won't remove the database, it will just shut off the services. If you want to do it programattically, the files you want are all in ~/.kde/share/config. You want kdedrc, nepomukserverrc, and nepomukstrigirc, they're ini-format files with square-bracketed headers and key-value pairs, you turn stuff off by setting it to false, of course. Where I work, we have a bunch of Debian/KDE workstations which NFS-mount user home directories -- desktop search is a network traffic disaster on the NFS server, so shutting this off is a pretty high priority for us. What we actually do is, all the clients have KDEDIRS set in /etc/profile, pointing to /usr/local/etc/kde, and in there, there's a $KDEDIRS/share/config directory with kdedrc, nepomukserverrc, and nepomukstrigirc with just enough settings to shut it off. In the KDE scheme, KDEDIRS entries override user settings, and the KDEDIRS is root-owned, so it's difficult for users to turn it back on. I think it's actually possible, though, if the users undefine KDEDIRS in their user-specific profiles, that might get around it. But, users also enjoy having the NFS server actually work, so we've been OK so far. Algol 68 specialist This is why I mentioned the config-file and system-wide ways to do it... -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120029.32996.rei...@bellatlantic.net
Re: duplicity backups are unidirectional? : local_directory - remote_host
[About remote back-ups w/ duplicity vs. rdiff-backup] I am unfamiliar with duplicity, but use rdiff-backup regularly, as well an rsync/rsnapshot combination. One thing that I like to do is to separate the remote transport from the actual back-up process -- there can be performance penalties for this, but it gives you better control. For instance, since you mentioned encryption, for your first scenario, you could make an sshfs mount of the remote filesystem on the back-up host, and then run duplicity (or another tool) in local-to-local mode. The performance penalty arises because the back-up tool now has to scan the source file system for changes over the network link, rather than running a daemon at the far end to scan it locally. Depending on the speed of your link and the size of your filesystem, this may or may not be a problem. I do this in production (with rdiff-backup, not duplicity) with a machine that actually backs up from a remote to a different remote -- the source remote is a read-only NFS mount, which therefore appears local, and the target remote is an iSCSI target on a storage appliance, which also appears local. The performance penalty is real, but it works for me. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201201252345.50383.rei...@bellatlantic.net
OT: Grepping style (was Re: Re (2): Laptop with external display running X11.)
On Fri, 02 Dec 2011 01:25:16 +0100, deloptes wrote: cat /var/log/Xorg.0.log | grep EE What's the purpose of this superfluous cat (and the pipe)? grep EE /var/log/Xorg.0.log I noticed that too -- I actually do this all the time, and I'm not 100% sure why. It's true that sometimes the input to grep is not a file, but the output of some script. I can rationalize that (thing-which-generates-data) | grep (pattern) will *always* work, and is thus preferred to grep (pattern) (thing), which only works on files, but I don't really know if that's the reason. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201112012240.36880.rei...@bellatlantic.net
Re: Full Disk Encryption
Hello, I am always interested in Full disk encryption for my laptop ( i5 + 3 GB ), but what makes me stop is the thinking of performance lag. Recently I have seen an ububtu laptop ( i5 + 4 GB ) with full disk encryption and it is performing normal, haven't found any lag... So I am interested to give the FUD a try on my own laptop. How can I proceed ? My laptop is debian wheezy with lots of important data.. so backup is must.. but what next ? What configuration will give me a better performance , LVM based or simple partition based ? Presently excluding swap I have 3 reiserfs partition for / ; /home and /movie ... no LVM. Like to hear some feedback from you guys.. I've had an LVM/luks-encrypted root partition (includes everything except /boot, on various logical volumes) for several years on two different Lenovo Thinkpads, and while I've never done any benchmarks, I haven't noticed any performance degradation at all. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20260911.14682.rei...@bellatlantic.net
Re: Latest iceweasel - package kept back
Dear All, An attempt to reinstall iceweasel produces: # apt-get --reinstall install iceweasel Reading package lists... Done Building dependency tree Reading state information... Done Some packages could not be installed. This may mean that you have requested an impossible situation or if you are using the unstable distribution that some required packages have not yet been created or been moved out of Incoming. The following information may help to resolve the situation: The following packages have unmet dependencies: iceweasel : Depends: xulrunner-8.0 (= 8.0-3~bpo60+1) but it is not going to be installed E: Broken packages So it appears that a new version of xulrunner needs to be installed. Where can I get this? Why is it not picked up automatically as a dependency? Any help will be appreciated. The relevant apt sources.list entries are: You need to specify the backport repo in order to get the dependencies to work, I think. Try apt-get install -t squeeze-backports iceweasel, which is very likely how you got it in the first place. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20232126.47960.rei...@bellatlantic.net
Re: Wifi with squeeze on Lenovo X61s laptop - difficulties
I've repacked linux-image-3.0.0-2-generic, wireless-crda and udev and have now discovered that I cannot install the udev package because it breaks package linux-base. So I cannot install the later kernel. It looks as though I shall have to forget using wifi on Debian squeeze. Has anybody got any ideas? There is a 2.6.39 kernel packaged in squeeze-backports, maybe it has a non-broken driver? -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20011907.05335.rei...@bellatlantic.net
Re: Hardware asset information
Hi, Is there a way to find total numbers of Hard disk attached or mounted on remote servers totaling around 200 Servers running debian Linux Server and also the number of RAM Chips attached to the system. Any utility or some gui tool from client desktop ? Hard drives are a bit tricky, because the system treats device-mapper devices and physical devices similarly in many contexts. I'd suggest scanning /proc/partitions for the devices that look like /dev/sdx (with no numerical suffix) or /dev/hdx, where x is a letter. For memory chips, you can do dmidecode and count up the populated type 17 entries, but beware, I believe it will also report un-populated DIMM slots. As for a desktop GUI tool, I have no idea. Probably there is, but that's not really my world. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201110032005.31708.rei...@bellatlantic.net
Re: daft newbie questions
Hi I've taken refuge from Kernel 3.0 and gnome3. can someone point me to where I can find libdvdcss and the w32/mpeg4 codecs.I can find reference to 32bit, but not amd64 I get these from debian-multimedia.org. And is firefox in a repo somewhere ?, ice weasle hasn't so opened as the lock file is set, is that in .mozilla ??? You don't need a new browser, you just need to clear the lock file. Look in your profile, in $HOME/.mozilla/firefox. Delete everything with lock in the name, but be careful, your bookmarks and browser state are in here too. If you don't value your bookmarks, you can just trash this whole directory and start over, but that's rather extreme, of course. For newer iceweasels, there are packages at mozilla.debian.net, although a word of warning, I am currently having a minor issue with the flash plug-in on iceweasel 6.0.2 on 64-bit squeeze -- occasional random crashes, it's usable, but annoying. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201110010737.44916.rei...@bellatlantic.net
Re: Worst Admin Mistake? was -- Re: /usr broken, will the machine reboot ?
I had a case where it had snowed, and instead of driving 50 miles in snow and ice with dodgy DC drivers, I'd work from home. Had my laptop, was doing work. Well, they scheduled a meeting for that afternoon (at about lunch time), so I got ready and headed in to the office. I typed halt in a window on my machine, and went to get my stuff together. Came back a few minutes later and found the laptop was still up. Had inadvertantly (I blame focus-follows-mouse) shut down a remote box, our production webserver... You can use molly-guard to protect against this -- installed on the remote system, it prompts for confirmation if a shutdown, reboot, halt, or poweroff command is entered in a remote shell. http://packages.debian.org/squeeze/molly-guard There's a legend that the name comes from an actual little girl named Molly, who was visiting the workplace and tried out the shiny red button. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201109142153.00984.rei...@bellatlantic.net
Re: Worst Admin Mistake? was -- Re: /usr broken, will the machine reboot ?
On Wed, Sep 7, 2011 at 1:00 PM, Bob Proulx b...@proulx.com wrote: jacques wrote: by error most of the binaries in /usr are erased (killing rm :-( Everyone has made that mistake at some point. I know I have! Not me! Though I did chmod -R /usr once. I noticed it immediately and cancelled. Most of the commands were broken, though luckily tar and scp still worked, so I copied over a backup and untarred it. This anecdote is brought up whenever anyone suggests skipping /usr /bin in backups is a good idea because the data doesn't change and would be recovered by OS reinstall anyway (yes I've heard that argument). Which brings me to another fun question. What's your worst administration mistake and how did you recover? I once tried to change the ownership of all the files in a user directory by doing something like chown -R newuser .* from within the directory -- I've forgotten what exactly I typed, but my motive was to get all the . files included in the scope of the command. Unfortunately, .* includes .., so the chown command hopped up to /home, and started switching the whole file system over to be owned by the new user. I caught it after it was taking a suspiciously long time, and after a minute or two, I figured out what had happened. Nothing focusses the mind quite like screwing up a live server. After a brief but intense look at the man-page, I figured out that what I wanted was chown -hR newuser path/to/target, and did that for both my initial task, and to fix the user directories that had been messed up. These days, I almost always use verbose options of commands, if they exist, so I can verify that they're operating in the expected scope. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201109131925.28318.rei...@bellatlantic.net
Re: System powers off randomly
Hello List, I'm running a Xen Host on Debian Stable. The architecture is amd64 and I'm using the standard Xen amd64 Kernel. The system is running on an Asrock E350M1 with the AMD e350 processor (dualcore, 1,6GHz) and two memory modules with 4GB each. The system was installed in April an was running fine. Then about two month ago from today a Kernel update was released, which I installed and rebooted the machine. From this point on the system crashes every now and than. This is almost certainly a hardware fault of some kind. One simple test you may be able to try, if your security policy permits it, is to revert to the kernel you were using earlier, and see if the problem goes away. My guess is that you will find the random shut-downs will continue, but there is no substitute for trying it. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201109111744.31758.rei...@bellatlantic.net
Re: kernel oops with Google Chrome after recent security update (2.6.32-35squeeze1)
On 2011-09-10, Camaleón noela...@gmail.com wrote: On Sat, 10 Sep 2011 09:19:52 +, Curt wrote: After installing a kernel security update yesterday, I'm getting random kernel oops when using Google Chrome. Anybody else seeing this phenomenon? Yes, a couple of users in the Spanish mailing list are also reporting this. Can you explain what is the current issue? I mean, does the oops happen when launching the browser? What's the content of the oops? I could not No, thus the word random. Appears to happen randomly (without predicability) while surfing, effectively crashing the browser. I am also seeing this, with Google Chrome Beta 14.0.835.159-r100066, and the most recent kernel update, 2.6.32-35squeeze1, on 32-bit. I am running some fairly flash-heavy stuff, and weirdly in one instance, the browser reported that a page had become unresponsive when the flash content on that page was still fine -- then the browser exited, and I saw the kernel messages. My kern.log entries are similar to those below, they begin with unable to handle kernel paging request at fff3 (I am on a 32-bit system). Anecdotally, it seems that it doesn't happen if the browser is the only thing running on the desktop, but I don't have a lot of data on that, I only did the kernel update a few hours ago. -- A. Sep 9 17:40:12 einstein kernel: [22866.749031] BUG: unable to handle kernel paging request at fff 3 Sep 9 17:40:12 einstein kernel: [22866.749043] IP: [8112f98d] m_stop+0x15/0x4c Sep 9 17:40:12 einstein kernel: [22866.749060] PGD 1003067 PUD 1004067 PMD 0 Sep 9 17:40:12 einstein kernel: [22866.749070] Oops: [#1] SMP Sep 9 17:40:12 einstein kernel: [22866.749076] last sysfs file: /sys/devices/system/cpu/cpu0/cpufreq/scal ing_cur_freq Sep 9 17:40:12 einstein kernel: [22866.749083] CPU 0 Sep 9 17:40:12 einstein kernel: [22866.749087] Modules linked in: powernow_k8 cpufreq_conservative cpufre q_stats cpufreq_powersave cpufreq_userspace parport_pc ppdev lp parport binfmt_misc fuse firewire_sbp2 loo p snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep snd_pcm_oss snd_mixer_oss snd_pcm snd_seq_mi di snd_rawmidi snd_seq_midi_event snd_seq snd_timer snd_seq_device acer_wmi snd rfkill led_class soundcore wmi psmouse i2c_nforce2 nvidia(P) k10temp pcspkr i2c_core edac_core evdev snd_page_alloc edac_mce_amd ser io_raw video output button processor ext3 jbd mbcache sg sd_mod sr_mod cdrom crc_t10dif usbhid ata_generic usb_storage hid ohci_hcd pata_amd ahci firewire_ohci firewire_core fan crc_itu_t libata ehci_hcd scsi_mod thermal thermal_sys forcedeth usbcore nls_base [last unloaded: scsi_wait_scan] Sep 9 17:40:12 einstein kernel: [22866.749200] Pid: 2413, comm: chrome Tainted: P 2.6.32-5-amd6 4 #1 Aspire X1301 (etc) (searching...) This can be related: http://www.google.vu/support/forum/p/Chrome/thread?tid=51f63d2301865cebh l=en Which fianlly points to this: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=640966 That's it. I guess the kernel team is working on it, God bless 'em. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201109101340.20782.rei...@bellatlantic.net
Re: kernel oops with Google Chrome after recent security update (2.6.32-35squeeze1)
On 2011-09-10, Stephen Allen marathon.duran...@gmail.com wrote: On Sat, Sep 10, 2011 at 09:19:52AM +, Curt wrote: After installing a kernel security update yesterday, I'm getting random kernel oops when using Google Chrome. Anybody else seeing this phenomenon? ---end quoted text--- No all is fine here on Wheezy 3.0.01 kernel latest Chrome Developer vs for Linux. I'm talking about the recent kernel security update in stable to: 2.6.32-35squeeze1 Just FYI, I checked the repo, and 2.6.32-25squeeze2 is now available. I'm installing it now, I don't know yet if it fixes the issue, but judging from the bug report referenced upthread, I think it's likely to be a fix. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201109101748.31722.rei...@bellatlantic.net
Re: Encrypted LVM container spanning drives?
I've never successfully set this up. Is there a way to encrypt two (or more) drives on a machine, then span it with LVM? I end up having to create one encrypted container per drive and having separate volume groups on each. I suspect I am looking at it wrong and that there is a way to do it. Its been a while since I've tried it, but what happens is that the first drive gets decrypted, but the second one doesn't, so the kernel panics because it can't find the volume group. Well, I've never actually tried it, but it seems to me that once you have your logical volume set up, there's nothing stopping you from doing luksFormat /dev/mapper/vgx-y and encrypting the thing. The underlying geometry of the LV shouldn't matter. The /etc/crypttab (and by extension the initramfs) should refer to the device by UUID, so as long as all the LVM stuff is in the initramfs, the kernel should be able to find it at boot-time. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201107291912.18773.rei...@bellatlantic.net
Re: clamav 0.97.1 not coming to squeeze-updates ?
On Mon, 27 Jun 2011 12:07:28 -0300, D G Teed wrote: On Mon, Jun 27, 2011 at 11:52 AM, Camaleón noela...@gmail.com wrote: Of course, squeeze-updates is the new volatile, nothing has changed. Our expectations for squeeze-updates to release clamav ahead of stable merely to be current are correct. Then it has to be a new policy. IIRC, not all of the clamav package updates reached the stable branch via volatile (now squeeze-updates), only those that closed security bugfixes. And I say this because I asked this same question here, months ago, and I was told so ;-) I don't know about policy, but my experience with clamav from lenny-volatile was that the engine itself actually did update, typically within a few weeks after it started squawking. My understanding for squeeze-updates was that it was intended to serve the same role for squeeze, i.e. that time-critical new versions which were not necessarily security-related would come through this pipe. ClamAV is the archetypical example for this, but given the recent announcements, I wonder if iceweasel might find itself in this channel also -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201106272244.26736.rei...@bellatlantic.net
Re: HELP: Number of CPU cores is not right
Hi all, Now I have one more problem is that the OS cannot recognize the right number of CPU Cores. The OS is Debian_5.0.8_amd64. The computer OS running has 48-core CPU, but the OS now only can recognize 32-core of CPU. The kernel version is 2.6.26-2-amd64. I tried to update the version of the kernel, but failed. Not sure what you mean here, but if you upgrade the whole business to squeeze, you will solve your problem. The issue is that the Debian stock kernel is compiled with a maximum CPU number of 32 in lenny, but 512 in 64-bit squeeze. Alternatively, you can rebuild your own kernel, starting with the default config file but changing CONFIG_NR_CPUS to 48 or more. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201106171901.14569.rei...@bellatlantic.net
Re: Weird Wifi problem -- WPA-EAP TTLS fails
On Fri, 20 May 2011 20:47:21 -0400 Andrew Reid rei...@bellatlantic.net wrote: So, apologies for the long-windedness, but what can cause EAP to fail? Do I need to add some libraries with more authentication schemes in them somehow? Obviously I have all the dependencies of wpa_supplicant, but is there something else? I don't know if I can be of much help, as I'm running EAP-TLS with FreeRADIUS, but you don't have any other takers yet. And all I can suggest is that you probably won't solve this without seeing the RADIUS logs, on what I assume is a Windows server, and I've no idea what they call RADIUS these days. It used to be IAS on Server 2003, and I've never had anything to do with that. The WAP itself is part of a Cisco Enterprise system. I'm not sure what the back-end authentication is, our workplace duplicates enterprise passwords across many authentication engines (to reduce password proliferation, a goal I heartily endorse). I do know that the Mac I used was not any kind of Windows domain member, and the Debian laptop also is not. I've put in a support query for the server-side logs, but the first-line support's response is it works on the Mac, our system is fine, Linux is not supported, and I have to admit that for a support team with scarce resources, that's not an absurd answer. I have asked them specifically for the authentication logs (and given them a precise time of the failed attempt and the originating MAC address), but haven't heard back on that yet. I've googled around a bit more since my initial post, and I'm starting to think I might actually be able to parse the wpa_supplicant logs, and maybe sharpen my question, possibly by figuring where in the EAP framework it's coming undone. What I suspect has happened is that the squeeze wpa_supplicant has some kind of new default that's breaking the process, and if I can just figure out what it is and set it to work like lenny did, I'll be fine. But, wpa_supplicant's option space is pretty big. I think I may be able to scare up another Linux laptop, and may even be able to get lenny on there, to try to close in on this. Anyways, thanks for your reply, mostly just thinking out loud here... -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201105211257.31395.rei...@bellatlantic.net
Weird Wifi problem -- WPA-EAP TTLS fails
Hi all --
I'm having a strange problem with my wireless connection, and I'm running
out of ideas.
I have a ThinkPad T510 laptop, running stock Debian squeeze. I use
the KDE desktop, but I think that's not an issue, because I've reproduced
the problem without any network managers or anything.
This machine has the Intel Centrino Wireless-N 1000 system
(PCI ID 8086:0084), which is supported by the stock iwlagn module.
It worked in lenny, with the backported kernel and drivers.
I can connect to my WPA-PSK access point at home, and to unencrypted
public Wifi systems, without any difficulties, but at work, we have a
WPA-EAP TTLS set-up, where it doesn't connect.
I can connect with my credentials on a colleague's Macbook, so my
account is evidently active, and the access point works, it looks like
the problem is on my system somewhere. Linux is unsupported at my
workplace, so I'm on my own.
The way I am connecting is in instructions all over the place:
ifconfig wlan0 up
wpa_supplicant -i wlan0 -c /etc/wpa_supplicant/work_wpa.conf
... which associates with the right SSID, then does this:
CTRL-EVENT-EAP-STARTED EAP authentication started
... then waits for maybe 30 seconds, then
CTRL-EVENT-DISCONNECTED - Disconnect event - remove keys
I have generated much more verbose logs with wpa_supplicant -dd etc,
but I really have no idea what I'm looking at. Some times it looks like it
times out, but I have some traces without timeout messages, which didn't
work. All of them have a line, EAP: Received EAP-Failure, which is the
thing that most looks like an actual solveable problem.
It seems that there can be a lot of variability in the log files,
for a while I was trying different options in the conf file, and seeing
if it looked like it was getting farther in the auth process. From that,
I did learn that it seems to do more with the right password than with
the wrong one, which suggests that *something* is working, but that's
about as much as I can get out of it.
My work_wpa.conf file is as recommended by my employer, and looks like:
ctrl_interface=/var/run/work_wpa
eapol_version=1
ap_scan=1
fast_reauth=1
network={
ssid=correct SSID, in quotes
key_mgmt=WPA-EAP
eap=TTLS
identity=correct username, in quotes
password=correct password, in quotes
anonymous_identity=anonym...@example.com
ca_cert=quoted/path/to/cert
priority=2
}
I have tried explicitly setting a phase2=autheap=MSCHAPV2, and
some others, and I've read about lots of other parameters.
I've found several related-looking posts by googling around,
which motivated me to try re-loading the iwlagn module with
swcrypto=1 and/or swcrypto50=1, but this does not change
the behavior.
So, apologies for the long-windedness, but what can cause EAP to
fail? Do I need to add some libraries with more authentication schemes
in them somehow? Obviously I have all the dependencies of wpa_supplicant,
but is there something else?
Thanks in advance.
-- A.
--
Andrew Reid / rei...@bellatlantic.net
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/201105202047.21826.rei...@bellatlantic.net
Re: Waking from the Dead
This just got harder. I'm trying to just do a reinstall but I only have a macbook to work from. And the installation media can only be a USB drive. I am having all kinds of trouble getting an ISO image onto the USB that will work. I can 'cat debian.iso /dev/disk1s1' well enough. And the machine will recognize the disk at start up, but it never sees it as a bootable device and just hangs. Many of the other instructions are assuming you have a working linux box, which I don't. The files/packages I need to download to build a bootable image I don't have and I can not get either -- apt-get is locked up on dependencies that I'm unable to resolve. Hi -- The instructions I used don't seem to have that requirement -- you just need dd, which Macs can do, I think. According to my (slightly dated, and possibly fragmentary) notes, I got boot.img.gz from: ftp://ftp.debian.org/debian/dists/stable/main/installer- i386/current/images/hd-media/boot.img.gz (NB 32-bit, and probably was Lenny when I did this, but others are supposed to work...) Then, plug in your USB device, do zcat /path/to/boot.img.gz /dev/sdx (substitute OS-specific nomenclature for /dev/sdx), mount /dev/sdx (it will have a bootable FAT32 file-system), and copy the ISO of your choice to the root of the device. Then unmount, and boot your installation target system from it. I have a dim recollection that the name of the ISO file mattered, and had to match a config entry on the USB device somewhere, but my notes, alas, don't cover that case. Also, the ISO can't be too big -- you'll want the net-install ISO for this. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201105142144.39055.rei...@bellatlantic.net
Re: Mystery on my laptop: KDE + nvidia + X + wheezy/sid
Hi, [ Lots of good info elided...] Any ideas where the problem might lie? I have a long shot -- I've had a bit of trouble with the nepomuk services on KDE4, although not NVidia-related (as far as I know), and not on experimental, as I generally stick with squeeze. Anyways, your traceback has nepomuk stuff in it, so as a first cut, I suggest disabling all that stuff. In the KDE gui, you can do this via System Settings - Advanced - Service Manager, and disable the Nepomuk Search Module. This may fail if the System Settings is one of the KDE apps that blows up, of course. In that case, maybe you can do something more blunt-instrument-like, like renaming the nepomuk executable or something. Otherwise, there's the usual is it plugged in stuff: I've sometimes had issues when Debian-packaged OpenGL updates clobber the NVidia-provided OpenGL library files, but have always been able to clear them by re-running the NVidia installer. Kernel updates can do this too, not sure why. (Your symtpoms don't really match this profile, but it's easy to try...) Also, the K desktop is vast and contains multitudes. If it's convenient, maybe create a fresh, blank account on the problem system, log into it in KDE, and let the config wizards do their thing, and see if the problem still manifests -- if it doesn't, then you've isolated the problem to one account's desktop settings. It's possible the 4.4 to 4.6 migration isn't as clean as it could be, and something's getting confused. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201105072202.47354.rei...@bellatlantic.net
Re: iSCSI + LUKS
Hi to all! I searched the web for finding an answer but with no luck. Anyone got an idea about how to first make iSCSI see a remote disk, then LUKS open the LUKS device and finally the filesystem on LUKS gets mounted. The problem now is that first LUKS noearly tries to create the /dev/mapper device but iSCSI has not run yet. Then iSCSI runs and the system sees the remote disk, and last, fstab tries to mount the filesystem, which of course fails because the LUKS /dev/mapper device is not there. How can I make the order: iSCSI, LUKS, fstab ? Presumably this is an issue for you at boot-time? I suggest mounting everything up and then re-building your initramfs -- it's pretty good at enabling the current configuration, and in particular, if you didn't have any crypto disks last time you built it, it may not have the crypto modules and/or /etc/crypttab may be set incorrectly. That's mostly guesswork though. One other possible catch is that you may want LUKS to refer to the device by UUID, if it doesn't already do so -- the device name can change from one iscsiadm log-in to the next, just as with USB and real SATA drives. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201104161401.49711.rei...@bellatlantic.net
Re: Asus A8N-E - debian-amd64 hangs during installation
I have box with sid i386 installed on it. It has two disks WDC WD1001FALS in raid5 array (md0) On md0 exists lvm. I tried to install debian-6.0.1a-amd64 but with no success :( I can confirm for you that Debian 6.0.1 will indeed run on this board, I was running it up until this past weekend, when I retired the system in question. I only ran 6.0 in 32-bit, but in the past, I've run 64-bit lenny on that board, again, with no issues. 1. Firstly I tried to install on prepared logical volume - kernel oops (dmesg can be seen http://strony.aster.pl/kjonca/instalka.amd64.log) 2. Then I tried to use some other spare disks; either sata or pata - no success also. Sometimes i can see kernel oops, sometimes I cannot even change console from installer. The only thing I can think of is pretty unlikely -- I have sometimes run into a situation where the installer will not run on a system, because of missing modules, but the full OS will run just fine. Since you apparently have a running sid, you could possibly set up the new system on a disk via debootstrap, then boot into it, and see if that works better. I think that's pretty unlikely, though -- my experience with this is that typically it's more common with an older installer and a newer motherboard. The A8N-E is from 2005, I'd expect it to be fully supported by the installer kernel. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201104052042.50284.rei...@bellatlantic.net
Re: RAID start at boot
Is there a way to make sure my RAID (level 1) won't be started degraded? On boot, one disk is found before the others, and the RAID is started before the others are seen. (They are seen at different times because I am transitioning from USB to eSATA, and the one eSATA disk is seen before any USB disk.) I start my RAID manually anyway (not as part of the boot process), so I'd be just as happy if it was never automatically started, but I *really* don't want it started with just one disk. According to the mdadm.conf man-page, you can specify an array with the name ignore in that file (and rebuild the initramfs, presumably), and this will cause mdadm to never automatically assemble the array. You could then presumably assemble it by hand specifying the name to mdadm -A whatever later on. I actually checked the man-page because I was *sure* there was a --no-degraded option in there somewhere. There is such an option for the mdadm command, but it's not clear if it can be gotten in to the boot-time environment or not. Also, I second the already-mentioned rootdelay idea -- that's probably better than ignore, but of course, it's your call. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110045.13142.rei...@bellatlantic.net
Re: Mdadm device reassignment
Hi all, I just installed the new stable Squeeze system, and moved over a raid5 from an older system. For some reason, it now wants to mount as md127 instead of md0. I've been unable to map it back to md0. [ Details elided...] root@jupiter:~# mdadm --assemble /dev/md0 /dev/sdb1 /dev/sdc1 /dev/sdd1 /dev/sde1 mdadm: /dev/md0 has been started with 4 drives. root@jupiter:~# cat /proc/mdstat Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10] md0 : active (auto-read-only) raid5 sde1[0] sdb1[3] sdd1[2] sdc1[1] 4395407808 blocks level 5, 64k chunk, algorithm 2 [4/4] [] unused devices: none This works fine, until reboot, at which point we're back to md127. In /var/log/syslog, I find this little oddity: Feb 17 15:33:35 jupiter mdadm[1580]: DeviceDisappeared event detected on md device /dev/md0 Feb 17 15:33:35 jupiter mdadm[1580]: NewArray event detected on md device /dev/md127 The whole thing has me pretty stumped, what am I missing? I have no explanation for how you got there, but the config getting re-set at reboot is probably because there's residual weirdness in the initramfs. Next time, after you fix it, try setting /etc/mdadm/mdadm.conf to reflect the desired config (I usually do this by putting the output of mdadm --query --detail --scan, leaving out the metadata=value bit), re-run update-initramfs, and then reboot. This may clear it, whatever it is. For the record, I've done a squeeze upgrade on a system with software RAID, and didn't have any issues, so it's not like it's doing it to everyone. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201102172052.52172.rei...@bellatlantic.net
Re: checking for changes in file size/permissions since installation?
Does apt/dpkg keep track of permissions and file sizes of the files which belong to a package? If so, how can this information be retrieved so as to compare to existing files on the file system? I looked into a similar issue a while ago, and as far as I can tell, apt does not do this. There are several utilities that will continuously monitor your system, and report changes in file sizes, permissions, etc -- they are host-based intrusion detection systems. Where I work, we use the Beltane/Samhain/Yule suite. Tripwire is also a good one, and is packaged for Debian. rant, severity=minor What I actually was looking for was a Debian-aware intrusion detection system -- I had a problem where, when I did package updates on all our workstations, the IDS would report all these file changes, and there didn't seem to be an alternative to manually OK-ing all of them, which is tedious and potentially error-prone -- if an attack occurs on update day, I am likely to miss it in all the spurious IDS traffic. It seemed to me that a sensible option would be to have an IDS that would notice when files had been changed by apt, and not report those changes, just fold them into the database of the system state. It's probably sufficient for my purposes to have a rule that says if the file has changed, but is controlled by a package, and changed within x seconds of that package being updated, update teh database to reflect this change, and do not report it. Obviously, the down-side of this is that adding any kind of do-not-report hook to the IDS is a potential exploit, since it could presumably be spoofed, but it seemed like a positive cost-benefit balance to me. I never did find such a tool. Some IDSs have a lot of hooks for custom scripts, so it may be possible to roll one's own, but I didn't get that far with it. /rant -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201102130831.55440.rei...@bellatlantic.net
Re: Help! Cannot mount XFS filesystem createt with Ubuntu on Debian
On Wednesday 09 February 2011 17:35:38 Federico Cislaghi wrote: Hello everyone, [ snippage ] Here is what I tried under Debian, based on the content of /etc/mtab under Ubuntu: root@atlantis:~# mount -t xfs -o rw,nosuid,nodev /dev/sda /mnt/multimedia/ mount: /dev/sda: can't read superblock root@atlantis:~# tail /var/log/messages Feb 9 23:31:23 atlantis kernel: [ 2290.237187] attempt to access beyond end of device Feb 9 23:31:23 atlantis kernel: [ 2290.237194] sda: rw=0, want=1953525168, limit=1953523055 Feb 9 23:31:23 atlantis kernel: [ 2290.237278] XFS: size check 2 failed The part of this that looks weird is that you said you partitioned the drive, with gparted, but you are apparently attempting to mount the device, /dev/sda, and not a particular partition, like /dev/sda1 or similar. Is this really what the Ubuntu system does? After the disk is attached to the Debian system, what's in /proc/partitions? What I find a little bit surprising is the output of fdisk: root@atlantis:~# fdisk -l /dev/sda Disk /dev/sda: 1000.2 GB, 1000203804160 bytes 255 heads, 63 sectors/track, 121601 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Sector size (logical/physical): 512 bytes / 4096 bytes I/O size (minimum/optimal): 4096 bytes / 4096 bytes Disk identifier: 0x Disk /dev/sda doesn't contain a valid partition table This part makes sense to me -- fdisk can't read gparted-style partition tables. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201102112003.37995.rei...@bellatlantic.net
Re: Build Deb install then move it
On Saturday 05 February 2011 23:01:56 Harry Putnam wrote: Once all that is working wouldn't I be able to move the whole thing onto the existing linux desktop by doing a network install of debian, formatting whatever space then use dd or maybe something more modern to plop the fleshed out OS from the virutal machine onto the bare install `/' disk. Or any other or better way to get this done. I don't know about better, but I can offer proven -- if you already have a partition on your system, it's pretty simple to just copy over a working installation. I do this to clone workstations from time to time. There are a small number of minor gotchas. Firstly, when copying from the working system to the destination drive, I use rsync with the --numeric-ids option, which ensures consistency between the destination's /etc/passwd and the actual UIDs. This is important for service accounts like ntp, which needs to own the /var/lib/ntp. Secondly, there are a few files that are sensitive to the switchover -- fix them up after you've done the copy. /etc/hostname and /etc/mailname may be wrong if there's a new hostname. /etc/network/interfaces might need a new static IP. The private and public host keys in /etc/ssh will be new, if you want the old ones, copy them over. If the host is Kerberized, you'll want to either copy or re-generate /etc/krb5.keytab. Finally, remove all the entries from /etc/udev/rules.d/70-persistent-net.rules, otherwise udev will notice the changed MAC address and generate a new interface name for it. In your case, the apache configs may need adjustments also, particularly ports.conf, which might need a new IP or hostname. When you copy a running system, there are a few messy bits, /var/run will have wrong data in it, and /var/log will have old-system logs, but in practice, I've found these don't matter -- modern init scripts automatically clean up /var/run, and the log stuff just carries on with the new info. The one thing this doesn't take care of is the boot-loader, you'll want to make sure your new disk has the one you want, and that the kernel and otehr config files are findable by it. Every few months, I also wonder if there's a slicker way to accomplish this, but I've never actually gotten around to figuring it out. I'll be following the thread to see if anyone else knows a good way. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201102060226.29110.rei...@bellatlantic.net
Re: alternatives for gcc
On Tuesday 01 February 2011 20:11:31 Joe Riel wrote: Why are there no alternatives, configurable with update-alternatives, for gcc? Seems like I should be able to configure whether /usr/bin/gcc is linked to gcc-4.3, gcc-4.4, etc. Of course I can just set the link manually (which I do), but ... Can't you set up new alterantives within the update-alternatives mechanism? I have never done this, but it was one of the options I was considering a while ago to control which version of an app was the default -- as it turned out, I was able to use the stow mechanism to good effect, because it was a non-packaged app, so I only got as far as reading the update-alterantives documentation. The man-page says that the --install option sets up a new batch of alternatives. I don't know if this will interact constructively with the package manager. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201102012101.13639.rei...@bellatlantic.net
Re: Can't reboot after power failure (RAID problem?)
On Monday 31 January 2011 10:51:04 dav...@alcor.concordia.ca wrote: I posted in a panic and left out a lot of details. I'm using Squeeze, and set up the system about a month ago, so there have been some upgrades. I wonder if maybe the kernel or Grub was upgraded and I neglected to install Grub again, but I would expect it to automatically be reinstalled on at least the first disk. If I remove either disk I get the same error message. I did look at /proc/cmdline. It shows the same uuid for the root device as in the menu, so that seems to prove it's an MD device that isn't ready since my boot and root partitions are each on MD devices. /proc/modules does show md_mod. What about the actual device? Does /dev/md/0 (or /dev/md0, or whatever) exist? If the module is loaded but the device does not exist, then it's possible there's a problem with your mdadm.conf file, and the initramfs doesn't have the array info in it, so it wasn't started. The easy way out is to boot from a rescue disk, fix the mdadm.conf file, rebuild the initramfs, and reboot. The Real Sysadmin way is to start the array by hand from inside the initramfs. You want mdadm -A /dev/md0 (or possibly mdadm -A -u your-uuid) to start it, and once it's up, ctrl-d out of the initramfs and hope. The part I don't remember is whether or not this creates the symlinks in /dev/disk that your root-fs-finder is looking for. It may be better to boot with break=premount to get into the initramfs in a more controlled state, instead of trying to fix it in the already-error-ed state, assuming you try the initramfs thing at all. And further assuming that the mdadm.conf file is the problem, which was pretty much guesswork on my part... -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201101312047.53519.rei...@bellatlantic.net
Re: module kvm-intel causing virtualbox problems
On Sunday 30 January 2011 18:15:56 bri...@aracnet.com wrote: I get an error message when I run virtual box and the inter-tubes led me to the answer of removing the kvm-intel module. I'm trying to figure out how it got loaded in the first place so I can keep it from being loaded, but I can't seem to track it down. Can some kind soul help me out. My intertubes told me that kvm is a virtualization module associated with qemu -- it may be that if you have the right sort of CPU, your kernel thoughtfully loaded it up for you, anticipating that you'd want the functionality. You may be able to prevent this by putting the module name in /etc/modprobe.d/blacklist. The comment at the start of that file describes the scope of the file, and distinguishes between hotplug events and autoloading, with this file controlling the former but not the latter. I confess I don't really understand this distinction. Whenever I have wanted to prevent a module from loading, adding the module to that file has always done the trick for me. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201101301938.12331.rei...@bellatlantic.net
Squeeze volatile?
Hi all -- I'm finally getting around to migrating some systems to squeeze, and I noticed that as of now there is no squeeze-volatile repository, although squeeze-backports exists. Does anyone know if there are plans for such a thing? Lenny-volatile was what allowed me to keep using the clam AV scanner without it going obsolete, so this is potentially important to me. Thanks. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201101271411.30097.rei...@bellatlantic.net
Re: Squeeze volatile?
On Thursday 27 January 2011 15:35:03 Andrew M.A. Cater wrote: On Thu, Jan 27, 2011 at 02:11:30PM -0500, Andrew Reid wrote: Hi all -- I'm finally getting around to migrating some systems to squeeze, and I noticed that as of now there is no squeeze-volatile repository, although squeeze-backports exists. I think you'll find there's now a squeeze-updates which supersedes and provides much of the functionality of the old volatile. Cool, thanks, that's what I was looking/hoping for. Also, I'm pleased to report that the lenny-squeeze upgrade went pretty smoothly. Looking forward to the release. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201101271703.31687.rei...@bellatlantic.net
Re: Weird server mystery: self-reset, mostly
On Monday 24 January 2011 01:47:27 will trillich wrote: Never seen this before -- all daemons and all user processes killed. Zap. It happened around 23:17 Chicago time (that's when the log-daemons quit logging). What would cause this? [ ... snippage ... ] *Anybody got a clue as to what might have happened to kill all daemons and user-processes in one swoop? This has been a rock-solid Debian server for years...* One thing that can take down a lot of daemons all at once is if the system thinks it's switching to run-level 1, or even shutting down -- it runs all those K-symlinks in /etc/rcX.d. Obviously this isn't *exactly* what happened to your system, since it showed 6 users from uptime, and your ps list shows init[2] (it's about 4/5 of the way down in your list), indicating run-level 2 (I think), but it may be related. Also 500+ days is impressive up-time -- you are clearly not doing kernel patches ... -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201101242128.31756.rei...@bellatlantic.net
Re: VirtualBox 4 fails building kernel module (the same with nvidia module)
On Saturday 22 January 2011 13:24:08 MRH wrote: Recently I updated VirtualBox from 3.2 to 4.0 (non OSE version). From that moment it stopped working, saying that the kernel driver (vboxdrv) is not loaded and suggesting to run /etc/init.d/vboxdrv setup This fails: -- Stopping VirtualBox kernel modules:done.. Uninstalling old VirtualBox DKMS kernel modules:done.. Trying to register the VirtualBox kernel modules using DKMS: Error! Bad return status for module build on kernel: 2.6.38-rc1.mrh.01 (x86_64) Consult the make.log in the build directory /var/lib/dkms/vboxhost/4.0.2/build/ for more information. Failed, trying without DKMS ... failed! Recompiling VirtualBox kernel modules: Look at /var/log/vbox-install.log to find out what went wrong ... failed! -- The above is the /var/log/vbox-install.log Below /var/lib/dkms/vboxhost/4.0.2/build/make.log -- DKMS make.log for vboxhost-4.0.2 for kernel 2.6.38-rc1.mrh.01 (x86_64) Sat Jan 22 18:07:34 GMT 2011 make: Entering directory `/usr/src/linux-2.6.38-rc1' LD /var/lib/dkms/vboxhost/4.0.2/build/built-in.o LD /var/lib/dkms/vboxhost/4.0.2/build/vboxdrv/built-in.o CC [M] /var/lib/dkms/vboxhost/4.0.2/build/vboxdrv/linux/SUPDrv-linux.o In file included from /var/lib/dkms/vboxhost/4.0.2/build/vboxdrv/include/VBox/types.h:30, from /var/lib/dkms/vboxhost/4.0.2/build/vboxdrv/linux/../SUPDrvInternal.h:35, from /var/lib/dkms/vboxhost/4.0.2/build/vboxdrv/linux/SUPDrv-linux.c:31: /var/lib/dkms/vboxhost/4.0.2/build/vboxdrv/include/iprt/types.h:105:31: error: linux/autoconf.h: No such fil e or directory make[2]: *** [/var/lib/dkms/vboxhost/4.0.2/build/vboxdrv/linux/SUPDrv-linux.o] Error 1 make[1]: *** [/var/lib/dkms/vboxhost/4.0.2/build/vboxdrv] Error 2 make: *** [_module_/var/lib/dkms/vboxhost/4.0.2/build] Error 2 make: Leaving directory `/usr/src/linux-2.6.38-rc1' -- I have home-build kernel, I had the same problem on 2.6.32 first, which worked nicely with VirtualBox 3.2 but stopped after installing VB 4.0 - unfortunately it seems it doesn't help going back to the VB 3.2 version, as the kernel driver module has been deleted and it fails building in the same way. So it seems there is a problem in building kernel modules. I have the same problem with nvidia proprietary driver - fails building. I suspect the problem is I build the kernel with --append-to-version: make-kpkg --initrd --revision=.mrh.01 --append-to-version=.mrh.01 but this helps me keeping things tidy. Now I have no idea where is the problem, is it a known bug or I do something wrong (but it worked to me until recently). And I'm afraid I have no idea which package is the culprit. For the home-built kernels, you want to also build the kernel-headers target, which will create the -headers .deb file, which you can also install. Use the same append-to-version as for the kernel image target, and it should work without further intervention. I have some really good notes on this somewhere, but can't find them just now, but as I recall, the make-kpkg docs are pretty good. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201101221826.41577.rei...@bellatlantic.net
Re: Exim4 : allow relaying for authenticateed users (LDAP,PAM over TLS)
On Sunday 09 January 2011 11:15:25 Frank Lin PIAT wrote: Does anyone knows a good howto. I am especially wondering how to instruct EXIM to use PAM/LDAP rather than the local /etc/exim4/passwd) It has been a long time since I did this, but my recollection is that Exim is one of those sensible applications that actually uses the system getent calls to figure out users and so forth, so as long as your /etc/nsswitch.conf file is set up to use the LDAP, Exim will automatically and seamlessly get it right. PAM don't enter into it. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201101091409.31356.rei...@bellatlantic.net
Re: set privoxy to rewrite http to https
On Saturday 01 January 2011 06:46:09 S Mathias wrote: I use the KB SSL Enforcer https://chrome.google.com/extensions/detail/flcpelgcagfhfoegekianiofphddcko f?hl=en so i could browse the net safer [i mean webserver - me] with using https, if available. Note that there is an EFF-authored Firefox extension that's supposed to have this functionality -- see here: http://www.eff.org/https-everywhere It works the same way that your privoxy works, by rewriting HTML pages on the fly according to a rule set, with the goal of putting https links on https-loaded pages, where https links are known to exist and work. You may be able to use this directly, if you're willing to switch to firefox, but even if you're not, it may be useful as a source of rules to apply to the tool you're using. I should probably disclaim that I don't actually use this extension, and that while I am a member of the EFF, I don't work for them, and am not affiliated with https-everywhere. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201101021429.00528.rei...@bellatlantic.net
Re: upgrade to squeeze fails for NFS-rooted system
On Sunday 28 November 2010 00:50:22 deb...@good-with-numbers.com wrote: Upgrading lenny to squeeze fails with the following message during apt-get upgrade: Yes, I do have an NFS mount--my whole filesystem: -- # mount /dev/nfs on / type nfs (rw) [...] -- So how do I upgrade? One option is to chroot into the exported file system on the NFS *server*, and manually do this part of the upgrade. You might need to bind-mount /proc to suppress some complaints. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201011281131.53985.rei...@bellatlantic.net
Re: Configuring RAID-1 boot partition
On Monday 22 November 2010 11:06:32 Peter Tenenbaum wrote:
I'm running squeeze on my desktop and recently decided to configure it for
RAID-1 as part of my recovery from a hard drive failure. I found an
article online about how to do this:
http://linuxconfig.org/Linux_Software_Raid_1_Setup . I followed the
article's recipe as best I could, but could not in general do exactly as
was shown for setting up the boot partition as the instructions seem to
correspond to GRUB 1 and I am running GRUB 2 (well, 1.98). When I
attempted to boot off of my RAID-1 boot partition, I got into grub and then
got the following message:
error: file not found!
Entering rescue mode...
grub rescue
At the moment I am booting off of a Debian Live DVD. Here is the content
of the grub.cfg file on the RAID boot partition:
set default=0
set timeout=5
menuentry 'Debian GNU/Linux, with Linux 2.6.32-5-amd64' {
set root=(hd0,1)
echo'Loading Linux 2.6.32-5-amd64 ...'
linux /vmlinuz-2.6.32-5-amd64 ro root=/dev/md3 quiet
echo'Loading initial ramdisk ...'
initrd /initrd.img-2.6.32-5-amd64
}
I'm unfamiliar with the details of grub2, so my first wish is that
you please share your results -- Grub 2 is supposed to be able to
read MD arrays.
Note that /dev/md3 is the RAID-1 array which is going to be root in the
fully booted system, and it is made of /dev/sda3 and /dev/sdb3. The boot
array is /dev/md1, which is made of /dev/sda1 and /dev/sdb1.
I think your set root is wrong, certainly one way, and probably
two. The first way is, grub counts disks from zero, so the first
partition of the first drive is (hd0,0), which is what you want, I
think. The probable second problem is that what you *really* want is
to point it at the RAID device. I don't know how grub2 names RAID
devices, but (hd) notation is probably not it.
Note that, because of the nature of RAID1, setting root=(hd0,0) will
probably allow your system to boot, but it's still not really right.
You should be able to type in the relevant bits in grub rescue mode,
you don't have to edit the files until after you've booted.
-- A.
--
Andrew Reid / rei...@bellatlantic.net
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/201011222041.45801.rei...@bellatlantic.net
Re: Configuring RAID-1 boot partition
On Monday 22 November 2010 11:06:32 Peter Tenenbaum wrote:
[stuff, then]
menuentry 'Debian GNU/Linux, with Linux 2.6.32-5-amd64' {
set root=(hd0,1)
echo'Loading Linux 2.6.32-5-amd64 ...'
linux /vmlinuz-2.6.32-5-amd64 ro root=/dev/md3 quiet
echo'Loading initial ramdisk ...'
initrd /initrd.img-2.6.32-5-amd64
}
I withdraw my earlier criticism of (hd0,1) -- it seems that Grub2 *does*
index partitions starting from 1, and not from 0, as legacy grub does.
Bad assumption on my part, sorry for the noise.
-- A.
--
Andrew Reid / rei...@bellatlantic.net
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/201011222115.02412.rei...@bellatlantic.net
Re: Debian stock kernel config -- CONFIG_NR_CPUS=32?
On Friday 22 October 2010 11:34:19 ow...@netptc.net wrote: In fact IIRC the additional overhead follows the square of the number of CPUs. I seem to recall this was called Amdahl's Law after Gene Amdahl of IBM (and later his own company) Either that's not it, or there's more than one Amdahl's law -- the oen I know is about diminishing returns from increasing effort to parallelize code. I don't know it in its pithy form, but the gist of it is that you can only parallelize *some* of your code, because all algorithms have a certain amount of set-up and tear-down overhead that's typically serial. Even if you perfectly parallelize the parallelizable part of the code, so it runs N times faster, your application as a whole will run something less than N times faster, and as N gets large, this serial offset contribution will come to dominate the execution time, at which point additional investments in parallelization are probably wasted. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201010222005.49579.rei...@bellatlantic.net
Re: Debian stock kernel config -- CONFIG_NR_CPUS=32?
On Friday 22 October 2010 03:22:02 Sven Joachim wrote: On 2010-10-22 03:15 +0200, Andrew Reid wrote: I recently deployed some new many-core servers at work, with 48 cores each (4x 12 core AMD 6174s), and ran into an issue where the stock Debian kernel is compiled with CONFIG_NR_CPUS=32, meaning it will only use the first 32 cores that it sees. For the record, CONFIG_NR_CPUS has been increased to 512 (the maximum supported upstream) in Squeeze. This is good to know. When I built my own 2.6.26, I also noticed that the maximum value offered by the configurator was 256 -- we'll likely be seeing systems with that many cores within a few years, if current trends continue. Basically, 32 is chosen a bit arbitrarily. But there are some problems with high values of CONFIG_NR_CPUS: - each supported CPU adds about eight kilobytes to the kernel image, wasting memory on most machines. - On Linux 2.6.28 (maybe all kernels prior to 2.6.29), module size blows up: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=516709. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201010222008.44192.rei...@bellatlantic.net
Debian stock kernel config -- CONFIG_NR_CPUS=32?
Hi all -- I recently deployed some new many-core servers at work, with 48 cores each (4x 12 core AMD 6174s), and ran into an issue where the stock Debian kernel is compiled with CONFIG_NR_CPUS=32, meaning it will only use the first 32 cores that it sees. For old Debian hands like me, this is an easy fix, I just built a new kernel configured for more cores, and it works just fine. But I'm curious if anyone on the list knows the rationale for distributing kernels with this set to 32. Is that just a reasonable number that's never been updated? Or is there some complication that arises after 32 cores, and should I be more careful about tuning other parameters? I googled around a bit, but didn't find much... -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201010212116.00023.rei...@bellatlantic.net
Re: Security policy
On Saturday 09 October 2010 15:34:58 Paweł Ch. wrote: Hi, I must create security policy for my company. Can someone send me example security policy? Especially with division to user, administrator and boss. There are a number of free public resources available from the US National Institute of Standards and Technology. A former employer of mine used Special Publication 800-53 as a baseline for a security policy. Besides providing a list of recommendations, it also has a pretty good discussions of the whys behind them, and the cost-benefit trade-offs that must be made. A list of NIST security division's publications is here: http://csrc.nist.gov/publications/PubsFL.html SP 800-53 itself is here, in PDF format: http://csrc.nist.gov/publications/nistpubs/800-53-Rev3/sp800-53-rev3-final_updated-errata_05-01-2010.pdf -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201010091642.51833.rei...@bellatlantic.net
Re: video card reccomendation
On Sunday 26 September 2010 10:46:38 Enrico Weigelt wrote: * Rick Pasotto r...@niof.net schrieb: I need to get an agp video card so take the video load off the regular bus. Also my new monitor has 1600x900 resolution and the openchrome driver doesn't recognize that resolution. I don't play games, so it's mostly for text, web, and an occasional video file. Don't buy Nvidia. They're totally reluctant against OSS, and their binary-only drivers are simply crap (dont want to flame on their bad coding practises here ;-o). Those companies IMHO don't deserve a single penny. I agree that nVidia's attitude towards open source is a problem. However, my experience with their binary drivers has been quite good. They're reasonably easy to install, and seem to perform quite well for me. In particular, they have one of the best distribution-agnostic binary-blob installers I've ever run across. As a Debian user, when I do have to resort to binary/proprietary software, having it only available as RPMs just adds salt to the wound. (Don't get me started about the older Intel C/C++ compiler installers...) -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201009262140.48466.rei...@bellatlantic.net
Re: Latest Lenny update: What new hardware is supported?
On Monday 06 September 2010 19:46:00 tw...@cstone.net wrote: I'm specifically interested in whether the Verizon USB 760 broadband wireless modem is now supported, but in general, when the kernel image is upgraded, where can I find a list of the new hardware inclusions without asking anyone to lead me by the hand? FYI, this device is supported in the backported 2.6.32 kernel. I run this on a ThinkPad of mine, for exactly this reason. Note that you will probably still have to *activate* the USB thing (i.e. bind it to a phone number) on a commercial OS, but you can *use* it in Linux. At least, that was my experience with the Virgin Broadband2Go service using the same device. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201009072047.20084.rei...@bellatlantic.net
Re: Straw poll: What browser do you use?
On Tuesday 07 September 2010 09:16:26 B. Alexander wrote: I'm just wondering, since firefox/iceweasel seems to be getting unusable. I have a 2.2GHz C2D box with an nvidia card at home, and a 3.0GHz C2D with a (lame) ATI card at work. I find that firefox (or xulrunner-stub) have memory leaks, and after a couple of days, it eats up a significant amount (10-30%) of memory. The work box has 3GB and the home box has 4GB. It also eats up a significant amount of CPU. Recently switched to Google Chrome on my netbook, and set up the Flashblock add-on -- Flash seems to be getting more common, and more of a CPU hog. The major downside is that the version I have seems unable to work with the PDF plug-in from Adobe, so I have to open them in an external app. It's annoying, but not a deal-breaker. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201009072056.11367.rei...@bellatlantic.net
Re: 802.11 WPA network key prompt despite GNOME keyring
On Saturday 14 August 2010 04:16:47 Stanisław Findeisen wrote: Hi I am trying to use a 802.11 wireless network that frequently disconnects me (low signal). The network is protected with WPA pre-shared key. The problem is I am being prompted for this key every time. Well, almost every. When I log in, I do not have to type the key. It is stored in the GNOME keyring (I use nm-editor 0.6.6 for that). But then, when it disconnects, I have to type it again and again. Why? And sometimes it doesn't prompt me, and failes to authenticate also. :-| I had to delete the network entry from the GNOME keyring (nm-editor), get *prompted* for the key, and *then* I was authenticated. 8-| What's wrong??! I've had a similar issue, and in my case, it seems to revolve around the network manager -- I can always clear it by doing sudo /etc/init.d/network-manager restart In my case, this is actually on Ubuntu 10.04, but the symptoms are identical. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201008142235.20924.rei...@bellatlantic.net
Re: amd64 does net detect my wired and wireless nic at installation
On Monday 19 July 2010 10:14:54 Paul Cartwright wrote: On Mon July 19 2010, Mathieu Malaterre wrote: On a recent DELL Opteron I had to use a squeeze installer (kernel 2.6.32) to detect my intel network card. I had a similar problem on my Dell laptop.. had to install ipw2200. http://wiki.debian.org/ipw2200 For those following along at home, I recently had a similar issue with drivers, but solved it differently -- the 2.6.32 kernel (same version number as Ubuntu 10.04) is in lenny-backports. Of course, if the issue is missing SATA or network drivers messing up your install, the backport kernel is not as useful, but still, it's an option. There was some talk at the time of the lenny release that a lennynhalf kernel release might be provided, similar to what was done for etchnhalf, but I haven't heard anything recently about that. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201007192122.01683.rei...@bellatlantic.net
Re: Current solution for ia32 on amd64?
On Monday 28 June 2010 12:51:28 ha...@softhome.net wrote: I am having trouble figuring out what the right way to install i386 packages on AMD64 Sid is. Specifically, I need to install some 32-bit dev packages so I can compile Wine. (for regression testing) I have ia32-libs installed, but it doesn't contain the development packages I need. The Wine wiki tells me to use ia32-apt-get. I do not have ia32-apt-get installed, and apt-get tells me there is no installation candidate. I can't find it on packages.debian.org I'm not sure if it will work for Wine, but I've started using virtualization for almost all of my 32-bit needs. There's a virtualbox-ose package in Debian lenny, it works, and it's almost trivial to set up a VM. The one place it doesn't work is for hosting a 32-bit browser on a 64-bit base system -- or rather, I haven't figured it out, since the VirtualBox documentation has some discussion of shared files. (I want my bookmarks preserved, and downloads to go somewhere where I can find/use them.) For regression testing of most things, it would be fine, but Wine is likely to require low-level access to the hardware, so it might not be as good for that, hence my earlier hesitation. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201006291546.45704.rei...@bellatlantic.net
Re: New amd64 system needs flash player
On Saturday 26 June 2010 14:32:40 John W Foster wrote: Thanks send it I,ll look at it. I have what is 'supposed' to be a pure 64 bit libflahplayer.so that I got from Ubuntu. It will not install. I tried just placing the lib in ~/lib64 but that did not work. I welcome any possible solutions. If you have it working where did you place the lib so that Iceweasel etc. could find it?? /usr/lib/iceweasel/plugins. I think the browser also searches /usr/lib/mozilla/plugins. My experience is with lenny systems. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201006261446.23112.rei...@bellatlantic.net
Re: Torrents killing my conection
On Sunday 20 June 2010 18:06:37 ABS Doug wrote: Downloading Ubuntu through Iceweasel went fine... thing is it went SO fast, I'm not sure it's really a good test. FYI, this effectively rules out the MTU issues I suggested earlier, so it was a useful test for that. The multi-OS character would seem to rule out hardware, so at this point the most likely candidate is network drivers. A bit of googling following up on your lspci output (showing an Atheros AR242x device) turned up this page: http://ubuntuguide.org/wiki/Ubuntu:Lucid#Atheros_Cards The upshot there seems to be that the Right Thing is to select the proprietary drivers option at Ubuntu-install-time, and things should work. If you didn't do that, you can follow the instructions on the page there to get recent madwifi drivers. If you already *did* select the proprietary drivers at Ubuntu-install time, and they're not working, then I suppose you could try the more recent madwifi drivers anyways... -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201006202026.57111.rei...@bellatlantic.net
Re: Torrents killing my connection
On Saturday 19 June 2010 02:02:17 ABS Doug wrote: I can't believe I'm still totally unable to figure out what is going on here. Ubuntu Netbook Edition 10.04 was unable to handle downloading torrents. [etc.] Is this only true of torrents, or is it also true of large downloads? What happens if you try to pull a few megs of something? I've had problems from time to time with MTU auto-discovery and TCP frame-size negotiations, and the symptoms are generally that small-packet, low-bandwidth traffic is fine, but big downloads stall out all the time. Modern routers (including WRT54Gs) are pretty good about this, but network drivers play a role too, and they do differ between the various OSs you've tried. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201006192121.34974.rei...@bellatlantic.net
Re: mdadm doing strange things
On Saturday 19 June 2010 14:20:27 Alan Chandler wrote: [ Details elided ] HOWEVER (the punch line). When this system booted, it was not the old reverted one but how it was before I started this cycle. In other words it looked as though the disk which I had failed and removed was being used If I did mdadm --detail /dev/md1 (or any of the other devices) it shows /dev/sdb as the only device on the raid pair. To sync up again I am having to add in the various /dev/sda partitions. SO THE QUESTION IS. What went wrong. How does a failed device end up being used to build the operational arrays, and the other devices end up not being included. My understanding of how mdadm re-arranges the array (including for failures, etc.) is that it writes metadata into the various partitions, so I agree with you that this is weird -- I would have expected the RAID array to come up with the sda devices as the only devices present. There are two things I can think of, neither quite right, but maybe they'll motivate someone else to figure it out: (1) Device naming can be tricky when you're unplugging drives. Maybe the devices now showing up as sdb actually are the original sda devices. Can you check UUIDs? This explanation also requires that you didn't actually revert the disk, you only thought you did, but then didn't catch it because the conjectural device-renaming convinced you that the RAID was being weird. (2) How did you revert the root partition? If you copied all the files, then I have nothing else to add. If you did dd between the partitions, however, you may have creamed the md metadata, and caused the system to think the sdb device was the good one. This explanation is unsatisfactory because, even if it's right, it only explains why that partition should be reversed, not the others, although if you didn't revert the others, they're copies, and you can't tell them apart anyways. Also, what happened to /etc/mdadm/mdadm.conf on the reverted root partition? Is it nonexistent on the one you're now booting from? There's potential for confusion there also, although I think the initramfs info will suffice until the next kernel update. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201006192115.16300.rei...@bellatlantic.net
Re: Moving /tmp to a separate partition. Advice?
On Sunday 23 May 2010 04:37:18 Klistvud wrote: Howdy, fellow Debianites! Given some extra hard drive space, I decided to move my /tmp dir (currently located under / ) to a partition of its own. I am looking forward to any advice, particularly of the been-there-done-that type: * how should I configure my fstab entry? How does Debian installer do it? Watch out for permissions -- /tmp is 1777 (rwxrwxrwt), it has to be world-writable and have the sticky bit set, which ensures that only users who create files in there can write to them. Permissions come from the mounted FS, not the mount point, so make sure you set these permissions while it's mounted. Because of the world-writability, security conscious admins mount it nodev and nosuid. If you're more careful, you can mount it noexec, too, but that will break some third-party software installers that work by examining your system, writing a custom config script inside /tmp somewhere, and then running it. So your fstab entry might look like: /dev/with/temp/ /tmp ext3 nosuid,nodev 0 2 * is there anything Debian-specific to watch for? Not that I recall. * is it true that setting /tmp permissions to non-executable, while hardening your box, prevents apt from working properly? Setting /tmp to non-executable by the noexec mount option does break things, but as I said above, my recollection is that it mostly breaks third-party stuff. I think the apt scripts are all in /var/lib/dkpg/info, and are run from there. Setting the *directory* noexec seems very bad, since the exec bit on directories controls the ability to cd to it, and turning that off would make it largely useless. As to why, on moderately-high-availability multi-user systems, I often put /tmp on a separate partition precisely so I can use mount options to globally control access. This is more important in a truly multi-user system than a home system, of course. Misbehaving apps rarely but sometimes blow the lid off of /tmp, and having it be on its own partition means this doesn't compromise the system as a whole, and you can easily figure out what's going on by seeing the logged errors and looking at df output. Some folks keep /var/log on a separate partition for similar reasons. Again, all of this is more important in a multi-user production environment. On my home systems, I mostly don't worry about this sort of thing. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201005231038.48482.rei...@bellatlantic.net
Re: Moving /tmp to a separate partition. Advice?
On Sunday 23 May 2010 18:46:29 Tom Furie wrote: On Sun, May 23, 2010 at 10:38:48AM -0400, Andrew Reid wrote: Setting the *directory* noexec seems very bad, since the exec bit on directories controls the ability to cd to it, and turning that off would make it largely useless. Just for the sake of argument *why* is setting /tmp rw- a bad thing? Surely if you put a file there, you know the full pathname, why would you need to list or search /tmp? Well, I don't actually know for sure that it's bad, but it seems to invite broken-ness. Recursive Makefiles are notorious for cd-ing all over the place, but then again, that's usually in the source tree, which may or may not be in /tmp, depending where you unpacked it. As a theoretical example, I can easily imagine an installer that might unpack a set of example configurations into /tmp, and then do an ls to grep out the one that matches the local output of uname -m to select it for further architecture-specific processing. I can easily imagine myself writing such a thing. So, I confess making it largely useless was hyperbolic, but I still think it's a bad idea. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201005231930.17311.rei...@bellatlantic.net
Re: executable won't execute
On Thursday 13 May 2010 20:36:41 Kent West wrote: I'm trying to start a daemon for Maple v 14, but when I try to run it, the system complains that the file doesn't exist. What's up? wes...@]goshen.acu.edu]:/usr/local/Maple_Network_Tools/FLEXlm/11.7: ls -lh total 2.8M lrwxrwxrwx 1 root staff6 2010-05-13 16:04 lmcksum - lmutil lrwxrwxrwx 1 root staff6 2010-05-13 16:04 lmdiag - lmutil lrwxrwxrwx 1 root staff6 2010-05-13 16:04 lmdown - lmutil -rwxr-xr-x 1 root staff 854K 2010-05-13 16:04 lmgrd lrwxrwxrwx 1 root staff6 2010-05-13 16:04 lmhostid - lmutil lrwxrwxrwx 1 root staff6 2010-05-13 16:04 lminstall - lmutil lrwxrwxrwx 1 root staff6 2010-05-13 16:04 lmremove - lmutil lrwxrwxrwx 1 root staff6 2010-05-13 16:04 lmreread - lmutil lrwxrwxrwx 1 root staff6 2010-05-13 16:04 lmstat - lmutil lrwxrwxrwx 1 root staff6 2010-05-13 16:04 lmswitchr - lmutil -rwxr-xr-x 1 root staff 979K 2010-05-13 16:04 lmutil lrwxrwxrwx 1 root staff6 2010-05-13 16:04 lmver - lmutil -rwxr-xr-x 1 root staff 962K 2010-05-13 16:04 maplelmg wes...@]goshen.acu.edu]:/usr/local/Maple_Network_Tools/FLEXlm/11.7: sudo ./lmgrd sudo: unable to execute ./lmgrd: No such file or directory wes...@]goshen.acu.edu]:/usr/local/Maple_Network_Tools/FLEXlm/11.7: ldd lmgrd /usr/bin/ldd: line 117: ./lmgrd: No such file or directory That ldd can't see it is just weird. Is the directory NFS-mounted? I have sometimes seen NFS mounts get confused and do stuff like this, with files showing up in some queries but not others. Changing the NFS export or mount options to more careful settings can help (e.g. use sync, turn off caching, etc.) Alternatively, I recall that if a file-system is mounted noexec, it can give confusing messages when you try to run things from it -- I think it's usually permission denied, but maybe my recollection is faulty? Other than that, given the evidence in the rest of the thread, I'm going to guess that it's some kind of file-system error. Check the logs (especially /var/log/kern.log) for errors, maybe fsck the filesystem, and try again... -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201005141917.57156.rei...@bellatlantic.net
Re: no alternatives for firefox/mozilla
On Thursday 06 May 2010 23:20:07 T o n g wrote: $ update-alternatives --display firefox update-alternatives: error: no alternatives for firefox. $ update-alternatives --display mozilla update-alternatives: error: no alternatives for mozilla. In the alternatives scheme, the browser is x-www-browser, so for example on my system: $ update-alternatives --display x-www-browser x-www-browser - status is manual. link currently points to /usr/bin/iceweasel /usr/bin/konqueror - priority 100 slave x-www-browser.1.gz: /usr/share/man/man1/konqueror.1.gz /usr/bin/iceweasel - priority 70 slave x-www-browser.1.gz: /usr/share/man/man1/iceweasel.1.gz /usr/bin/google-chrome - priority 35 Current `best' version is /usr/bin/konqueror. The x-www-browser is respected by K desktop apps, and in my config, they all now use Iceweasel instead of Konqueror when I click on links. I don't know what other apps might or might not respect the x-www-browser setting, but I would think any desktop app packaged for Debian should do it. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201005071106.48303.rei...@bellatlantic.net
Re: usb raid1 pendrive boot
On Friday 07 May 2010 08:22:00 deloptes wrote: Jozsi Vadkan wrote: Did anyone managed to boot from RAID1? Lenny gives this error message: mdadm: no devices found for /dev/md0 You'll need a small i.e. 20-30MB partition out of raid to boot from with i.e. initrd which loads the md driver and your raid device becomes visible. Just for the record, it actually *is* possible to boot off a partition that's in RAID1, but you might have to use grub. The reason it works is because grub can read past the md-raid1 metadata without getting confused -- since the raid is not running when grub is looking at the disks, it only sees one of the two, but since they're mirrored, and since grub doesn't *write* to the partition, everything works. I have several production systems that do this. It's a bit complicated to set up, but generally, it goes something like: - Install a basic system onto one drive, with the correct size of partition (call this /dev/sda1). - Set up an incomplete RAID1 array, with one device and one missing, with the second drive -- this is /dev/md0, consisting of /dev/sdb1 and an empty slot. - Configure /etc/mdadm/mdadm.conf, and re-build your initramfs so that it has the md modules and knows about the array. - Mount /dev/md0 somewhere and copy all the data off /dev/sda1 to it. It's very important to get the *new* initramfs onto the array. - Install grub in the bootloader of the 2nd drive. (The installer already did this for the first one). This actually isn't essential, but you may need it if/when /dev/sda fails. - Boot on to the incomplete RAID array -- root=/dev/md0. - Add /dev/sda1 to the array, and sync. Now you have a bootable software raid-1 system. I'm not saying this solves the OP's problem, necessarily, but it's perfectly possible. I've never tried it with LILO or grub2, but I would think it would work. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201005071303.20994.rei...@bellatlantic.net
Re: Password messed up
On Friday 07 May 2010 21:15:17 Don wrote: Thanks, and if anyone has more ideas on fixing my password/login problem, please help! Coming in late to this thread, but if your K desktop is OK, then you actually have this problem quite well isolated to something in the KDM log-in process. It's not X, and it's not your desktop itself. By the time KDM has switched over to you, it's writing errors into a file, .xession-errors, in your home directory -- you can check this for clues to what's going on. Another option is to blitz kdm's config files, and re-configure or re-install the package. I'm not familiar with how sid does this, but you can do dpkg -L kdm, and look for files in the /etc directory -- for the current Debian stable it's /etc/kde3/kdm. I suggest re-naming the analogous directory (don't delete it, you might want it back), and then doing apt-get install --reinstall kdm, or whatever the analogous operation is in aptitude or synaptic, if that's your preference, which should re-build the configuration section. That's likely to fix it, at the cost of un-doing any customizations you've done. If blitzing and re-building the kdm config fails, then it's probably some kind of interaction between kdm and something else... -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201005072325.22749.rei...@bellatlantic.net
Re: xen
On Friday 16 April 2010 22:53:48 Mike Viau wrote: Sat, 17 Apr 2010 01:55:32 +0800 john...@hotmail.com wrote: Hello there, Ny name is Ian, from Perth Australia. I have a question to be answered please. I'm a linux user, so I use suse, debian, ubuntu and gentoo, and fedora.I'm interested in exploring the capabilities of xen and have done so an all the above.I'm in a brand new instance of squeeze and have just perused the packages in synaptic for the first time.I have recently used karmic lenny including a xen live cd which uses lenny as a dom host.This has lead to submitting bug reports because the collection of packages that are used to support virt-manager and xen are extensive and quite problem ridden. I am now to try the latest packages provided by squeeze. I can't help but notice THERE IS NO XEN KERNEL. I just did a search on packages.debian.org for xen, and came up with these: http://packages.debian.org/sid/xen-hypervisor-3.4-amd64 http://packages.debian.org/sid/linux-image-2.6.32-4-xen-amd64 This appears to be the Xen hypervisor and a Dom0-enabled kernel. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201004170919.17265.rei...@bellatlantic.net
Re: NIS user member of plugdev, gnome-mount of flash drive raises error
On Thursday 25 March 2010 18:38:36 Joseph Lenox wrote: I'm running lenny (5.0.4); and trying to get USB flash drive mounting in a way that doesn't involve hand-adding every user to the plugdev group (we're running NIS). I tried the pam_group approach, and id says the user is in the plugdev group, but I'm still getting a permissions error from DBus. Adding the exact user to the plugdev group on the local machine worked as far as the mounter. An alternative approach is to open up permissions on the device, by finding the right MODE=0660 line in /etc/udev/rules.d and changing it to MODE=0666. Then the group membership won't matter. I don't actually know how to do this, but there are promising-sounding entries in /etc/udev/rules.d/91-permissions.rules, you might try editing the usbfs-like devices section there as a starting point. It's also possible that the gnome-mount is setting some permissions somehow, you may need to edit that also. Sorry to be so vague, it's just another approach. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201003251912.56820.rei...@bellatlantic.net
Re: Transferring files over SSH in the console
On Sunday 21 March 2010 18:52:36 Dotan Cohen wrote: I am managing a small embedded device that I SSH into over the LAN. To run commands, I use KDE Konsole, and to transfer files I use Konqueror and SFTP. I understand that SFTP also runs over SSH, so is there a way to send files in Konsole as well? I am familiar with the FTP commands such as cd, lcd, put, and get. Are there equivalent commands for SSH terminal connections? As others have commented, you can use scp or sftp. However, I can imagine that embedded might not have the sftp service or the scp executable. If that's your case, you can always do: # cat file | ssh remote 'cat destinaton' i.e. pipe the file through a simple SSH invocation of cat on the remote system. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201003212314.32985.rei...@bellatlantic.net
Re: Transferring files over SSH in the console
On Sunday 21 March 2010 23:14:32 Andrew Reid wrote: On Sunday 21 March 2010 18:52:36 Dotan Cohen wrote: I am managing a small embedded device that I SSH into over the LAN. To run commands, I use KDE Konsole, and to transfer files I use Konqueror and SFTP. I understand that SFTP also runs over SSH, so is there a way to send files in Konsole as well? I am familiar with the FTP commands such as cd, lcd, put, and get. Are there equivalent commands for SSH terminal connections? As others have commented, you can use scp or sftp. However, I can imagine that embedded might not have the sftp service or the scp executable. If that's your case, you can always do: # cat file | ssh remote 'cat destinaton' Pardon my replying to myself, but I've now seen a bunch of the rest of the thread, and it seems to me that, if the set of commands you want to do is repeatable, then maybe what you want is to do most of the remote-system operations through SSH commands this way? Someone else may have already suggested this, but something like: # cat file.tgz | ssh remote 'cat dest.tgz' (or scp, if available) # ssh remote 'tar -xf dest.tgz' # ssh remote 'sh dest/installer' (or whatever) # ssh remote 'cat dest/install-log' remote-install-log (or whatever) This way, you still only have the one shell, and/but you pay the price in having to prefix all the remote operations with ssh remote. However, you could script this on the local system (which is, I think, why you want a single session, right, so you can script it?), and then the extra typing doesn't really cost you much. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201003212325.49742.rei...@bellatlantic.net
Re: (OT) gnash vs. flash (was Re: Why does installing gnome ...)
On Wednesday 17 March 2010 20:15:51 Neal Hogan wrote: On Wed, Mar 17, 2010 at 4:44 PM, Ron Johnson ron.l.john...@cox.net wrote: Enough DDs are sufficiently practical for there to be a non-free tree, and Christian Marillat does yeoman's work with http://www.debian-multimedia.org/. no flames intended Ok . . . that's fine, but I'm still curious about the number of debian-user@ folks who appreciate or abide by or whatever the Stallman-point-o-view. P.S. -- . . . and if I'm lucky, why Stallman might not be relevant to the linux project. The Stallman-purist position actually has a significant practical upside -- it's impossible to pirate free-as-in-freedom software, which means an admin can give users a free hand on the systems they work with, and there is almost no danger of my employer being embarassed by some kind of license violation if they talk about it on their blog or redistribute it or something. This makes the sysadmin's job easier. Steering around the legal issues in commercial software isn't difficult, usually it boils down to don't redistribute, but it can be tricky to police. That said, I am not myself such a purist -- I have lots of non-free packages on my various systems. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201003172115.48914.rei...@bellatlantic.net
Re: aptitude package filtering
On Sunday 14 March 2010 17:54:25 Freeman wrote: On Sun, Mar 14, 2010 at 05:23:11PM -0400, Yavuz Yetim wrote: Hi, I would like to list all packages installed from testing. How can I do that? First limiting to ~installed and then searching ~archive(testing) gives me packages that are... I think you'd be better off with something like $ apt-show-versions | gerp testing I love this list -- I've been mucking about with Debian for years, and had never run across apt-show-versions. I even discovered that there's an update for my lenny-backports version of OpenOffice. On my system, packages show up with the distro name, not the distro status (i.e. squeeze and lenny, not testing or stable), but aside from that it works as described. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201003141819.37071.rei...@bellatlantic.net
Re: Disabling Swap Activation Message
On Thursday 04 March 2010 22:12:59 Tech Geek wrote: I installed Debian Lenny on my PC (/dev/hda1) without any swap partition. I recompiled my kernel to boot without using an initrd. Now when I boot to my base system, I still see following messages from the init scripts on my console: Activating swap...Done . Activating swapfile sawp...Done After doing a find it seems that these messages are spitted out by scripts mountall.sh and checkroot.sh files in /etc/init.d/ directory. My question is what is the best way to disable these messages and the associated function with those messages? Should I just comment out the relevant swap portition in those two init files? Seems cleaner to me to either boot with a noswap kernel option, or edit /lib/init/vars.sh to always set NOSWAP to yes. Both mountall and checkroot seem to check that variable. Of course, I haven't actually tried it, so it might not even work. If it does, my personal preference would be to change the default boot options in the bootloader config -- I use grub, and it's already customized to have the right root= line, so adding noswap to the kopt line would be natural, and would be preserved automatically across kernel updates. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201003042234.03224.rei...@bellatlantic.net
Re: apt-proxy doesn't work
On Sunday 21 February 2010 20:17:56 Gonzalo Gorosito wrote: Hi guys, I just installed the apt-proxy and I can't get it working. Using the config file as it comes here's my log file: [ Good stuff snipped... ] Any clue? I recently switched to approx from apt-proxy for this sort of thing, I found it much easier to deal with. YMMV, of course -- not sure what I was doing wrong with apt-proxy, my symptoms varied from yours. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201002212040.02325.rei...@bellatlantic.net
Re: Basic Statistical Tool
On Monday 15 February 2010 13:00:36 Ogya Chief wrote: Dear All, I need a basic statistical tool such as the Data Analysis Add-in in MS Excel. I have installed MS Office 2007 on Debian Sid but the add-in installation fails. I know Gnumeric has similar functions to Excel's but it lacks some features for graphing that I need. For example, in Excel, linear regression produces graphs in addition to the table of data. I need this feature. OpenOffice.org Calc does not have data analysis feature at all. I have also looked at R but it has a steep learning curve. I need a tool to do an assignment this week. Are there any tools similar to the Excel Data Analysis tool? What happened to KOffice in Sid? apt-cache search koffice does not turn up anything. Any help will be very much appreciated. I'm unfamiliar with the Excel Data Analysis tool, so I may be misunderstanding the requirement, but have you looked at Gnu Octave? It's very similar to MatLab in its interface, and has quite powerful analysis tools, although I don't know how good the statistical suite is. It certainly has a friendlier learning curve than R. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201002151322.19640.rei...@bellatlantic.net
Re: unionfs
On Sunday 07 February 2010 17:32:45 Alex Samad wrote: Hi I am looking at trying out unionfs, I notice there is onlythe unionfs-fuse package - is this the way debian is going with unionfs or how do you use the kernel version - can't seem to find any comparisons between the 2 The Debian-packaged kernel overlay file system is aufs, which seem to be a fork and/or successor of unionfs. I have had some difficulties with overlay file systems in lenny, I have a diskless cluster where I wanted to overlay two different NFS-mounted file-systems, and was not able to solve the problem within the scope of lenny-packaged tools. Aufs does work in general, but not in the case where I wanted it. I'm not aware of any comparisons between unionfs and aufs, but my intuition is that the kernel system should have better performance. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Installing FreeBSD under Xen on Debian Lenny
Hi all -- I'm attempting to install a FreeBSD domU HVM under Xen using the stock Xen packages in Debian lenny (Xen version 3.2), and I'm running into a problem trying to boot the FreeBSD install CD on the VM. I'm getting a BTX halted error, which seems to indicate a misconfigured BIOS, but I'm getting nowhere beyond that with Google. The dom0 is Debian lenny amd64, on an ASUS P5Q Pro motherboard, which has hardware support for Intel virtualization tech. (Enabling this is the most commonly reported solution to the BTX error I've seen, but it was eanbled out of the box for me.) The CPU is an Intel Core2 Quad Q9400, which also has the necessary hardware support, so I don't think there's a hardware problem. The scheme I am trying to use is to have the VM boot off an image of the install CD, and then run the installer inside the VM to get everything set up. I have done this successfully for a Debian HVM domU, although there were some networking issues which I think are unrelated. The FreeBSD CD image I'm using boots successfully in VirtualBox, so I know it can at least get that far. Has anyone done this successfully? Can you share your VM config file and/or provide clues? I can of course provide arbitrarily terrifying levels of additional detail. Thanks in advance. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: USB disk shows up late at boot
On Saturday 16 January 2010 12:33:32 Hugo Vanwoerkom wrote: This has been solved by Ben Hutchings and was reported as http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=534324 The solution was to put ums-cypress in /etc/initramfs/modules and rerun update-initramfs -u for that kernel. Linux debian 2.6.32-trunk-686 Excellent news. Also, thanks for posting the solution to the list, where others can find it. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Slow NFS
On Saturday 02 January 2010 10:55:18 Chris Davies wrote: I'm curious whether anyone has good a decent read/write speed using NFS (v3 or 4) between a number of Debian testing based systems. I've got one box exporting a number of filesystems using NFS v3, and three others mounting various combinations of those file systems. I see less than 10% disk throughput on the remote boxes that on the local one. Switched gigabit LAN throughout. I'd expect some hit, but 90% degredation seems awfully high. There is a performance-tuning section in the NFS Howto, with several tips, including simple tests for measuring performance. http://nfs.sourceforge.net/nfs-howto/ar01s05.html My own experience with NFS on a server with several tens of clients is that the biggest determiner of performance is the type of file operations you're doing. Anything that requires many seek operations, like grepping for complicated regular expressions, or writing data character-by-character, will do very poorly. Things that operate sequentially, or in very large chunks, do much better. In general, copying is fast, searching is slow. I think NFSv4 is much the same, although it may introduce a few more options, and so be more tunable. I have no direct experience with it. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: debugging initrd image
On Saturday 26 December 2009 15:11:59 Hugo Vanwoerkom wrote: Hi, What is the way to debug an initrd file and drop into a shell? You can use the break options (detailed in another reply) to get at it live during the boot, but I have often found it useful to just unpack the thing on another system and look around. The initrd image is a gzipped cpio archive. man cpio for more details on how to unpack it. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: LAME for Lenny
On Thursday 24 December 2009 14:19:12 Bernard wrote: Hi to Everyone, On my older Sarge system, I used 'lame' to encode wav files to mp3. However I can't find 'lame' for Lenny. Could someone tell me where to find the appropriate package ? This package is in the debian-multimedia repositories -- a bit of googling should find you a nearby mirror, which you can add to your sources.list. It's been a while since I was current on this, but I believe they're excluded from the main repos due to patent and/or license encumbrances. If you care about such things, you might wish to encode to a more open format, like ogg. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: USB disk shows up late at boot
On Tuesday 22 December 2009 07:48:05 Hugo Vanwoerkom wrote: Andrew Reid wrote: It's likely the devices aren't being recognized in the initramfs -- possibly they require kernel modules which are not present by default. But I would think that to be the case of the custom kernel, not the Debian kernel. The small custom kernel recognizes both USB drives, the Debian kernel recognizes only one of them. There's no obvious reason (to me) why it should go one way or the other. Both custom kernels and the initramfs want to be small, so they can both be selective. I can easily imagine a scenario where, for convenience, you include all USB mass-storage drivers in your custom kernel, but the initramfs decides USB might not be needed at boot-time, and leaves a few out. I don't actually know, one way or the other, I'm really just guessing at this part. If you know which modules drive these devices, add them by name to /etc/initramfs-tools/modules (one module per line, I think), and re-generate your initramfs with update-initramfs. This should allow the udev scan in the initramfs to see the devices, and set them up earlier. I have no idea, except to compare both initrd images. One option is to compare the dmesg lines from the two kernels, they generally report what they're doing during the device scan. Another option is to query the live device when it's plugged in to the system, and ask it what module is being used. I don't actually know how to do this, but I'm pretty sure it's possible. There are lots of udev tools, maybe one of those. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: USB disk shows up late at boot
On Monday 21 December 2009 15:52:29 Hugo Vanwoerkom wrote: Hi, I have 2 internal ATA HDD's and 2 disks in external USB enclosures. When you boot (this is Sid) the 2 USB disks report their presence between the messages: 'Loading, please wait...' and 'Init 2.86 booting' in the very beginning of the boot process. Now the funny part: in my homegrown kernel both show up together. But with recent Debian kernel images only one shows up. I have a delay of 10 secs. in initramfs-tools but that makes no difference. The 2nd USB disk shows up eventually, but after 'Init 2.86 booting' when it is too late to be of use by fstab. This isn't the first time I've asked this, but nobody seems to have an answer. It's likely the devices aren't being recognized in the initramfs -- possibly they require kernel modules which are not present by default. If you know which modules drive these devices, add them by name to /etc/initramfs-tools/modules (one module per line, I think), and re-generate your initramfs with update-initramfs. This should allow the udev scan in the initramfs to see the devices, and set them up earlier. Init 2.86 booting is a very important milestone in the boot process, it marks the transition from initramfs activity to root file-system activity. Anything you want to do *before* that has to be in the initramfs. Or, as the other responder mentioned, you can just stick with a custom kernel. I used to do that, but I like getting security updates. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Serious clock problem
On Wednesday 04 November 2009 17:53:03 FTALOVER wrote: Hi all, I'm running VMW ESXi 3i (3.5.0) on an 4 CPUs DELL Poweredge 1950. This machine is hosting 8 virtual machines, 5 running Etch and 3 running Lenny. All guests were running fine until I upgraded the kernel to linux-image-2.6.26-2-686_2.6.26-19 in the 3 boxes running Lenny. Now those boxes are repeatedly freezing because the system clock seems to stall. Forcing the kernel to use acpi_pm as clocksource instead of TSC and disabling TSC doesn't seems to help. So, does anybody have any suggestion on how to fix this? I've got a Dell Latitude D630 that does this in lenny, I suspect it's a kernel bug, since the hardware was stable under etch. I got some relief by using clocksource=acpi_pm noapic as kernel arguments at boot-time, but it's done it one time since then -- sounds like you tried something similar. I can't find the bug-report thread I was looking at just now, but if you google the error messages, you can find a bit of mailing-list traffic (mostly Ubuntu), and some discussion of work-arounds, mostly involving kernel arguments. Sorry this isn't more helpful... -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Help, the template debian/rules is too simple
On Monday 02 November 2009 23:42:54 waixy zhou wrote: I admit what I'm doing is a little beyond my current skill, after all, I'm just a user of Debian not a developer. But I think it is better using apt/dpkg to manage the software/tools I need in my research than just using ./configure make make install. I prefer to use the ready-made packages if there is any. I googled, but unfortunately I can't find any. Do you know about stow? It's kind of a poor-man's package management system. It's pretty sophisticated, but the vanilla use-case for it is, you build applications inside a directory, /usr/local/stow/app-name, and the installer creates bin, lib, share, and so forth inside that directory. Then you run stow app-name from /usr/local/stow, and it creates sym-links so that your software shows up under /usr/local/bin and /usr/local/lib and so forth, just as if you had built and installed it with --prefix=/usr/local. The beauty of this, of course, is that you can *un*-stow the thing, when you upgrade or when you're finished with it, and it cleanly removes it without leaving piles of cruft in /usr/local. This might be a more appropriate solution to your problem, if your software doesn't have complicated dependencies. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Security.debian.org confused?
On Sunday 01 November 2009 05:36:00 Florian Weimer wrote: * Andrew Reid: http://security.debian.org//srv/security-master.debian.org/ftp/pool/updat es/main/e/expat/libexpat1_2.0.1-4+lenny1_amd64.deb This should have been fixed by now. During an internal migration, incorrect package metadata was pushed to the security mirror network. Sorry about that. Fix confirmed, it works now. Thanks! -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: a cautionary tale w/ successful recovery
On Sunday 01 November 2009 17:57:49 Andrew Sackville-West wrote: So here is the real success part of the story: my backups worked! I had weekly backups of /etc and daily backups of /home. Since I'd not done any work of consequence in about 24 hours, I had not lost data! Restoring was a simple matter of copying over from the backup server, fixing up a couple of permissions and moving on. Lessons learned: 1) don't do risky things in DOS 6.22 when tired... you can't trust DOS to behave in a consistent manner (maybe) 2) keep a copy of the boot sector lying around (on another machine!!) 3) keep a copy of dpkg --get-selections lying around Congratulations! If it were within my power, I'd award you the Order of the Clue, the one with the *nice* ribbon. For the sysems I back up at work, we do the dpkg --get-selections thing, but I've never kept a copy of the boot sector -- that's an excellent idea. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Security.debian.org confused?
Hi all -- I'm having trouble with what seems like it should be a routine security update on lenny -- I did apt-get update; apt-get upgrade as per usual, and got this: mec:~# apt-get upgrade Reading package lists... Done Building dependency tree Reading state information... Done The following packages will be upgraded: libexpat1 libmozjs1d xulrunner-1.9 3 upgraded, 0 newly installed, 0 to remove and 0 not upgraded. Need to get 8233kB of archives. After this operation, 28.7kB disk space will be freed. Do you want to continue [Y/n]? Err http://security.debian.org lenny/updates/main libexpat1 2.0.1-4+lenny1 404 Not Found [IP: 128.101.240.212 80] Failed to fetch http://security.debian.org//srv/security-master.debian.org/ftp/pool/updates/main/e/expat/libexpat1_2.0.1-4+lenny1_amd64.deb 404 Not Found [IP: 128.101.240.212 80] ... E: Unable to fetch some archives, maybe run apt-get update or try with --fix-missing? The elipses conceal the fact that all three files had the not-found problem. I tried re-updating and --fix-missing, no joy. Doing host security.debian.org gets this: mec:~# host security.debian.org security.debian.org has address 149.20.20.6 security.debian.org has address 128.31.0.36 security.debian.org has address 128.101.240.212 security.debian.org has IPv6 address 2001:4f8:8:36::6 security.debian.org mail is handled by 10 klecker.debian.org. The fetch URL looks wrong to me, and manually trying it in the browser seems to confirm that it's mangled. However, I can manually go to http://security.debian.org/pool/updates/main, and the stuff is in there. I haven't changed my configuration, but I'm willing to believe I missed a memo and have drifted out of date. Anyone got any clues? -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Cluster FS ?
On Thursday 22 October 2009 21:17:05 Brent Clark wrote: Hiya Would anyone be able to make any recommendations for a Cluster Filesystem. Not a recommendation, exactly, but a thought -- it is possible to do third-party implementations, you apparently don't have to buy it from IBM. A colleague of mine is attempting to do this on a CentOS system. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: debian whitelist, posts that never show up and long delays
On Saturday 17 October 2009 11:49:36 Nick Lidakis wrote: I seem to be having some trouble with my posts on this list lately. I have no trouble sending mail to my gmail account or myself at my verizon.net address; they show up instantaneously. My Veriozn ISP, in a move to combat spam, has recently required users to send mail to port 587 instead of 25. I am a new mutt user using Debian sid with ssmtp. In my muttrc I have: set sendmail=/usr/sbin/ssmtp -au nlida...@verizon.net -p 587 -ap $password I am also a Verizon customer, and my (possibly mangled) understanding of the situation is: - If you're using Verizon's mail servers as your SMTP hosts, you can use port 25, and/but you must authenticate, and you must use a verizon.net or bellatlantic.net from-address. - If you are using a third-party mail server outside of the verizon.net domain, and using SMTP to connect, then port 25 is blocked, and you must use port 587 to connect to these systems. Allowed from addresses and authentication requirements depend on the server. - If you're using a web-mail client (e.g. for gmail or yahoo or whatever), then, again, no configuration changes are required, since all the SMTP action is outside of Verizon's network. You don't say what mail relay your ssmtp command is configured to use, but if it's outgoing.verizon.net, then you can probably continue to use port 25 without difficulty, since you clearly have a Verizon address. I use kmail on Debian lenny, and send mail via authenticated ESMTP on port 25 to outgoing.verizon.net, and it's working fine. It's remotely possible I'm misconfigured and will be sorry when the switch-over hits, but I read the instructions carefully, I think I'm OK. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: why can't i use grub with an ext3 /boot partition?
On Sunday 13 September 2009 18:58:17 Robert P. J. Day wrote: ok, so let's move beyond anecdotes and get into conjecture -- does anyone know the debian installer well enough to know *what* circumstances will cause said installer to refuse to offer grub as a bootloader, and only present lilo? A potentially useful experiment is to try installing grub manually on the partition, and see if it works, or if it doesn't, what errors it reports. The installer may be suppressing error reports. I believe you can view the installation log during the install, from an alt console, but I'm not 100% sure it's possible, and I certainly don't remember how. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: why can't i use grub with an ext3 /boot partition?
On Sunday 13 September 2009 03:33:50 Robert P. J. Day wrote: possibly related to an earlier post that discussed grub but, when installing lenny, if i choose to create a separate, primary partition for /boot (a long-time habit), if i select ext3 for that partition, i am not offered the chance to use grub as a bootloader. as best i can tell, that /boot partition must be both: I conjecture that you are doing Something Else wrong. I have many machines that have ext3 /boot partitions, including several that were set up initially as lenny machines. I also have this habit, and I generally intend to use ext2, but sometimes I forget to switch it from the installer's filesystem default, which is ext3. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [Debian] Re: Grub boot problems
On Monday 24 August 2009 22:46:55 Glen Lee Edwards wrote: Emanoil Kotsev wrote: Glen Lee Edwards wrote: kernel /boot/vmlinuz-2.6.26-2-686 root=/dev/sdb1 ro quiet you have the root already root(hd1,0) is used for setup, I think. The only reason I posed the question is because all my Dell PCs have this problem, and every time Debian upgrades the kernel APT redoes munu.lst, so the computers won't boot. It doesn't take long to fix the problem, but it would seem that a better solution rather than constantly having to modify menu.lst would be to fix the problem once and for all on the PC end. What is groot set to in your menu.lst? It's possible/likely that just setting the right groot and re-running update-grub will fix things. If leaving it out works, then (hd0,0) is a likely candidate for the right answer. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: help for kernel panic
On Monday 24 August 2009 11:12:49 Alan Greenberger wrote: On 2009-08-22, Andrew Reid rei...@bellatlantic.net wrote: That would seem to narrow it down to a corrupt initramfs, or, as you already suggested, motherboard hardware issues. Thanks for confirming the md5sum and the suggestion. It also fails boot to level 1. I used rescue to turn on /etc/default/bootlogd but nothing gets logged to /var/log so I guess it doesn't get that far. I would have liked to know where it fails, since it does boot rescue CDs. I am giving up assuming it is the motherboard. That's probably a good bet, but if you're feeling adventuresome/insane, there's still that initramfs angle to explore. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Strange log-rotate problem
On Saturday 22 August 2009 02:09:09 Sven Joachim wrote: On 2009-08-22 01:08 +0200, Andrew Reid wrote: My problem is that I can't figure out who is rotating /var/log/auth.log. It's currently being rotated every day, and retained for a week. If you are using sysklogd (the standard syslog daemon in Etch), the answer is that it uses savelog, not logrotate. See bug #44523¹. The good news is that in Lenny and later, rsyslog² is the standard syslog daemon, and it uses logrotate. Of course, upgrading an existing system will not change your syslog daemon. Thanks, this is helpful. I found the sysklogd cron entries, but I further thought the daily one (in /etc/cron.daily/sysklogd) wasn't rotating auth.log -- it uses syslogd-listfiles to get the set of files to rotate daily, and when I ran it interactively, it returned an empty string. However, as a sanity check, I instrumented /etc/cron.daily/sysklogd, and sure enough, it *is* doing the mystery rotation. So, now I can adjust it to comply with policy, which solves the immediate problem. The remaining mystery is, why does the syslogd-listfiles give different answers interactively versus inside a cron script? Probably some environment thing. There is also an Etch backport of rsyslog, if you would like to use it without upgrading to Lenny. I'll probably be upgrading to Lenny in a few weeks anyways, so I'll just wait, I think. However, I did want to mention that I am a big fan of backports, they've helped me out a lot over the years. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Strange log-rotate problem
Hi all -- This is a long shot, but I thought I'd ask here. I have a multiply-upgraded Debian etch box, which is also a log host (i.e. many other hosts log on it via the UDP port that syslog provides), and our policies regarding log retention have recently changed. My problem is that I can't figure out who is rotating /var/log/auth.log. It's currently being rotated every day, and retained for a week. I spent a lot of quality time today with the logrotate documentation, and I'm confident that it's not in any logrotate scripts. I also checked out the cron-driven log rotation that's done by the scripts that come with the sysklogd package, and that package seems to be set up to rotate it weekly -- this may be working, but never getting the chance, because the daily rotations are colliding with the weekly effort. I know that rsyslogd provides logrotate packages, many of my systems work that way, but this system does not have rsyslogd installed. There are also some residual syslog-ng scripts, but they don't appear to be active. What I *do* know is that whatever is rotating the auth logs is cron-triggered -- they all have 06:25 time-stamps, suggesting they're run from /etc/cron.daily somewhere. But I've looked at all those scripts, and none of them seem to do it. Possibly relevant is that this system is very old, and has been transplanted to new hardware several times -- I think it started out as Debian potato, and has been steadily upgraded over the years, so it could be left-over functionality from some ancient package that's mucking things up. So, my specific question is, is there anything *else* besides logrotate or sysklogd scripts that can do log rotations? Some obscure cron thing that doesn't show up when I grep for auth or log, because it's doing some kind of crazy pattern-matching thing? Thanks in advance... -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: help for kernel panic
On Friday 21 August 2009 18:11:27 Alan Greenberger wrote: I have a system with Lenny installed from the KDE installer CD. It was working fine for half a year. Powering it on after being off for two weeks, it starts to load Lenny then dies with: Failed to execute /init can't open auto Kernel panic - not syncing: Attempted to kill init! iA linux-image-2.6.26-1-686 Version 2.6.26-13 The md5sum of /boot/vmlinuz-2.6.26-1-686 is 824cfba2eac12d0c09747c0bd3426e4e I think your checksum is OK, with a caveat -- it's the right checksum for the advertised 2.6.26-13 version of that package, which is distinct from 2.6.26-13lenny2, the latter being the most recent 2.6.26-1 kernel for Debian lenny, which in turn is not the most 2.6.26 kernel, there is a 2.6.26-2 out. I checksummed it by pulling the file out of /var/cache/apt/archives, manually unpacking it in a working directory, and running md5sum on the resulting $DIR/boot/vmlinuz-2.6.26-1 file. So, you're apparently behind on kernel updates, but your kernel does not seem to be corrupt. That would seem to narrow it down to a corrupt initramfs, or, as you already suggested, motherboard hardware issues. I don't have a huge amount of experience with this, but I did once have a similar issue -- I had a server that wouldn't fully boot, it would just hang, always in different places in the boot sequence. But, it could boot to single-user mode, and if you then started all the /etc/init.d services manually, it would run fine for months at a time. I never did figure that one out, I eventually got rid of the machine. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Unable to install mencoder debian package from debian-multimedia.org.
On Tuesday 11 August 2009 17:58:48 Thomas Anderson wrote: I have the following entry in my sources.list file: deb http://www.debian-multimedia.org lenny main But I cannot install any package from that repository. Try a mirror, there's a list of them at: http://debian-multimedia.org/debian-m.php I've got an Intel quadcore CPU (which means I'm on the AMD64 architecture, right?) and am running Debian stable (Lenny). Yes, those chips are amd64 architecture. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Etch to 5.0.2 upgrade failed - Encrypted filesystem will not boot
On Tuesday 11 August 2009 17:41:48 line...@ruiner.halo.nu wrote: Ok I guess the system is just hosed. If no one has any more suggestions in the next couple days I will reinstall. I will never trust Debian upgrades again, at least not when encrypted filesystems are in use. Well, all I can say is that it worked for me. It's pretty clearly an initramfs problem, since it works for your other kernel. It's also very weird (as you've remarked) that you can apparently initialize the encrypted partition via luksOpen from within the initramfs, but then not mount it -- I'm assuming you checked all the obvious things, like whether or not your candidate mount point (/a in your example) existed. I have one more nontrivial suggestion -- I suggest installing the 2.6.24 etchnhalf kernel. You'll have to pull it from the etch repositories. It's possible that running a new kernel install and corresponding update-grub and so forth will either (a) work, or (b) give a more meaningful error message. Also, please understand that I mean no disrespect, but I feel compelled to remind you of some possible stupid-mistake solutions: - Does your 2.6.26 kernel have the same boot options in menu.lst as the one that works? Are they the default kopt options, so they get propagated to new kernels by update-grub? If you manually added encryption after your etch install, they might not be. - Is the menu.lst that's modified by the package manager the same one that the boot-loader is actually using? I once had a system that had somehow gotten both /grub and /boot/grub directories, both with menu.lst files in them, only one of which mattered. - Along similar lines, is update-initramfs writing its files to the correct place so they're read at boot time? That's about all I can think of. Good luck. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Etch to 5.0.2 upgrade failed - Encrypted filesystem will not boot
On Wednesday 05 August 2009 19:54:50 line...@ruiner.halo.nu wrote: I tried configuring fstab to use the UUID from blkid, but I had the same problem. Could the problem be that the SCSI drives are not coming up until cryptsetup has loaded? Hi again lineman (and list). Just for another data-point, I have just now finished upgrading my laptop, with the dm-crypt-encrypted hard drive, from etch to lenny. There were several minor issues with packages, but the crypto part worked fine. (Anticipating doing this myself is part of why I was following this thread...) I *do* see the Driver 'sd' needs updating message, but my system boots, so that's apparently unrelated. My /etc/fstab file doesn't use UUIDs, it lists the device-mapper names of all the devices, and it works. This shouldn't matter in any case for the root FS, it's mounted before /etc/fstab is read, and the device name is taken from the kernel args. I tried booting with the break boot option, which drops you into the busybox shell at the init-premount step, and tried to see if I could manually set up the crypto, but it's a bit complicated, and relies on environment variables which are set for the scripts, but apparently not set in the shell. One thing you *can* do easily is, boot with the break option, and from within the resulting shell, run /scripts/init-premount/udev, which will create all the devices. You can then do an ls in /dev, and see if the relevant hard drive partition (/dev/sda5, in your case) is are present -- this tests the udev step pretty directly. Hope this helps. -- A. -- Andrew Reid / rei...@bellatlantic.net -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
