aRTee

Yeah sure, you just filled it,.... :D Glad it now works for you. I am curious what exactly is wrong with the standard kernel Lirc drivers...

From what lsmod tells, the driver is there, so that's not the problem. From what I gather, pixelview is not a kernel driver - so you won't find it there. It's a userspace driver, and I'm not sure if I get it correctly, but the output of lircd --driver=? actually confirms that it is available, so just put it in the conf file /etc/sysconfig/lircd and it should be used. Check with ps -ef | grep lirc if it is indeed used. What happens if you do: service lircd stop irrecord /dev/lirc/0 and then follow instructions? Or if you do: cat /dev/lirc/0 and press some buttons? Are you sure your dongle didn't break in the meantime?

:D No, I was aiming at the comment of scarecrow "or Mandy decided that this is not free anymore..." But you're right, due to the Intel packages, for instance proftpd had to be left out. Not a big issue IMHO, but Mandriva guys were not open about that...

Solarian, isn't that just a hint to drop MSWin USB scanner: mine works fine, as do all my other usb periferals: webcam, mice, cardreaders, usb connected digital cam....

I see the serial module in my system: ll /lib/modules/2.6.12-12mdk/kernel/3rdparty/lirc/drivers/lirc_serial/ total 8 -rw-r--r-- 1 root root 6687 Sep 9 18:16 lirc_serial.ko.gz So I don't know what your issue is, but it's not this. What happens when you do (as root) : lsmod are any lirc modules present? My desktop output: lsmod | grep lirc lirc_i2c 8036 1 lirc_dev 11812 1 lirc_i2c i2c_core 17360 12 cx88xx,rivatv,lirc_i2c,asb100,i2c_sensor,i2c_viapro,tuner,tvaudio,msp3400,bttv,i 2c_algo_bit,tveeprom Your /etc/sysconfig/lircd looks good to me... What's your output of ls -l /dev/lirc/0 On my mdv05le system (which also has a homebrew receiver), I have nothing in my /etc/modprobe.conf, and my /etc/sysconfig/lircd is practically identical to yours, minus that the device is /dev/lirc0 but other than that, no difference...

About the Intel stuff - you did notice the add during install, right? Anyway, there is now an 'interactive firewall'. More info here: http://qa.mandriva.com/twiki/bin/view/Main...ractiveFirewall Why does everyone always suspect that Mandriva makes things 'no longer free' (as in beer) if it goes missing? When have they ever done that? (Maybe Iwas asleep at the time?) This is one of my main reasons to support Mandrake/Mandriva: all they do is GPL. So only the external code (which they include as a service to paying members) is not open. As the saying goes: never ascribe to malice what can be explained perfectly by incompetence. That applies regularly to Mandriva management and communications. Not in this case though.. :D

You can get the boot.iso (read my site, I used it for my latest installations) and install from ftp (or http, nfs, local hd). So you don't actually have to wait. Since you want to download the iso images I gather you do have the bandwidth. You could just get the mirror contents (~10GB) and install from local hd / local network.

johnh123, it depends also what you call ati cards - I have a Rage Mobility, it works fine. lavaeolus, it's ok if you want to use a supported kernel; I actually think that it doesn't matter much, but it's your call. The bad part about that is that you most likely won't have suspend to disk working. What you could also do is use kernel mm for the time being, until a security update comes out for the stock kernel, then switch to that one. The graphics cards do have quite a bit to do with various suspend modes, in that they are often used with their proprietary drivers, and from the docs from ATI at least I gather that they don't find the suspend stuff interesting/important enough to work at it; besides, there are plenty of other things missing. But lavaeolus is correct in that acpi is problematic. Not so much in the sense that it's new - it's been in the linux kernel for over 2 years (since before 2.6 I think). The real problem is with broken implementations that the manufacturers then fix in the windows driver - so people don't know things are patched. Most commonly messed up are DSDT tables and such. And that can be problematic to fix, for popular notebook models it is usually doable to find info on how to fix them but not often without recompiling the kernel... I'm waiting for the day that one can find Linux preloaded fully functional laptops that actually have a sticker that says: built for Linux. I'm not holding my breath though... ;) lavaeolus, you only recently stumbled upon this site? Man where else do you get your fix?? :D

You're worried about the kernel on a laptop not getting security updates? Hmm,... wouldn't worry too much. In fact, unless it's a security hole that can be exploited if the firewall is up, I usually don't bother. Of course, my laptop and desktop are behind my FW... I'd say go ahead, try the mm kernel - it may well work fine for you, it does for me. I think most of the Lucent winmodems work - if I recall correctly, I have one of these, and in the past I have made it work; haven't bothered for Mdv05le and now Mdv06 since I've never had a use for it. BTW Welcome to Mandrivausers! and most of all: enjoy Linux :)

I wouldn't worry about the high cpu load - I have also deinstalled kat on my machine - the related high loads were from kded and mhonarc, but kat caused those to do whatever they were doing. They are working on solutions, I guess with a 120GB /home partition mostly used by one user, kat has a bit much on its hands to catalog... Just deinstall it and you'll be fine. On the other hand, if you're happy with your current system, just keep it and update the latest stuff that came out, new FF, new OOo2 and then some, and you'll be fine. BTW, my uptime: That's from my server, the old machine I had, before I got this one and loaded it with Mdv05LE, was running for over 450 days on Mdk9.2. Three vnc + icewm sessions, no X local. Some claimed that Mdk was not stable, has only recently gotten good, etc. Nonsense. 450 days is really nothing to laugh about; I used to joke to colleagues who used 'some other platform' that no-one on that platform could have the uptime I had - some nasty trojans and worms had come up in Feb 04, and again in April, so they had all had to reboot... Well, not that it matters, but anyone telling around that Mdk wasn't stable just had no clue. So I agree if you're happy, don't change, but you will not have a lesser experience if you do change; just typing urpme kat is all it needed for me.

Jon, the reason to use it is that the main weak point of urpmi it not the tool itself, but the dependency on the ftp mirrors. I recently had an error when trying to use urpmi that said: "that's not a nice response from the ftp server". I had to try again a few times, and suddenly, it did work. Smart remedies this weakness by allowing parallel ftp mirrors for the same packages/repositories; read the features page: http://labix.org/smart/features Seems to me like a real improvement - one cannot (apparently) expect mirrors to always function. That said, I haven't tried it yet.

arctic, I think you normally won't get sig warnings at installation, could you try one or some of the following: imwheel, lirc, setserial, librrdtool2 (dependency of lm_sensors), gqview, openoffice.org-go-ooo-2.0-0.m129, openoffice.org-go-ooo-kde-2.0-0.m129.3mdk, libSTLport4-4.6.2-1mdk (dependency of OOo2pre) This is a list of packages that I got signature warnings with, and they are what I consider useful software, so in case you don't know these programs, give them a try ;)

Thanks arctic. Chris / anna, good to know they will fix it, however, I'm seeing these with the powerpack dvd, not just external ftp/web repositories, so this is not going to be fixed for mdv2006... supposedly this dvd is the same thing that will be the boxed product, hence my 'paranoia'. Any others?? John, did I stroke your feathers the wrong way? Can't say that it was intentional. Hope you get over it at some point. This has NOTHING to do with working as root. I don't see why you want to drag bvc into this, firstly how can you be so sure he agrees with you, secondly, are you afraid you need someone on your side because your reasoning is flaky and you ran out of arguments and hope he may have some for you? As for your lack of counter arguments, I take it that you either agree with mine or haven't processed them properly yet. I got one more for you: From the manpage of rpm: From www.seifried.org/lasg/software/: From this it is clear that the signature is used to check both the integrity and origin of packages. So go ahead, go on believing you know all that. IMO your position is based on belief, not knowledge. Fine with me, freedom of religion and all. Now be my guest and show us some more sarcasm, who knows, some may find amusement in it. In case you care to rebut my arguments, I might follow up. BTW any others who see the warnings when installing stuff from the official Mandriva iso's? PS: AVG works like a charm! Maybe now I can sleep at night once more. :lol:

This is about proper behaviour. Just as the root example, bvc can ignore proper behaviour and he'll be fine. Some others will be in trouble that they can't get out of at some point. Do we want only those people who have inside knowledge about Linux to be using it without issue? Are those on MSWIn who have a spambot without knowing it at fault, those who run one-click email trojans (anna_kournikova_naked.jpg.vbs), or could the system possibly be set up in a smarter way that prevents issues? If the answer to the latter is yes (when would anyone EVER have to execute an attachment received via email? Just make those attachments non-executable and it's done - except that MSWin doesn't allow this due to the way it works), then the system is at fault and the user is not the only one to carry the blame. Sure, the user should/could be educated about how to use his computer - but that's just not realistic. In UNIX, the proper behaviour is not to run as root. Specialists like bvc can ignore that, but that doesn't make it proper behaviour, it just proves that it _is_ doable. As for signatures and source, AFAIK the signature is signed on top of the whole package, just as a regular email pgp signature, The signature tells you that the package is from someone who has the other end of the key and confirms package integrity. If I'm mistaken about this, I take back what I said here. I have done some googling and find that indeed the signature check is pgp based and is the only means that the end user has to verify the origin of the package. The man page also mentions this. I agree that at this moment it's not a big issue to ignore the warning, I'm just saying that this is an indication that things aren't working as they should. And that is just some 'unpolished' behaviour from what I believe to be the best distribution from Mandriva/Mandrake so far, only possibly outdone or met by the most recent releases of other distributions (K-Ubuntu 5.10? SUSE10? Still have to try). Besides, if Mandriva is the first distribution whose users get trojans on their systems (which may not be far away if ignoring / clicking away the warning is the standard way to install any package), it's really bad for Linux in general and Mandriva in particular. So basically, I'm not speading paranoia, but I see a gaping whole that can easily be closed, with just a bit of effort - that should go with building packages on the developers side, and with responsability on the users side. Just suppose that all packages are signed properly. As was the case with Mdv05LE for most packages. Imagine an ftp server gets hacked, people who get infected with trojans because of that _are at fault themselves_ ! They chose to ignore the "incorrect signature" warning. Now, if most packages give the "incorrect signature" warning when one installs them, and people get their machines infected with trojans after some ftp server gets 0wn3d, you can't possibly blame them; not ignoring those sig warnings would have left them with an unusable barebone system. So the blame would be on the packagers and ultimatively on the distribution maker, in our case Mandriva, because in fact, the mechanism to avoid trojans is already broken. The reason I started this topic is to know if others also see this, because as I just said, mdv05le was very clean in this respect. I'm still not sure if something went wrong in my installations (though it happens on all 3 machines I installed)... I would like to know what others think of this - maybe it's time for a poll... In any case, to rebut some of your arguments that I disagree with and to comment further: So based on _what_ do you actually believe that a package is truly from a certain source? The only thing that one may assume is that it comes from the ftp server that holds your repository - and even that could be hacked around, purely technically spoken. Moreover, YOU may be able to judge (I still don't see based on what), but how can novice judge anything? It will just scare and worry them. I think you base your believe that it's ok to continue on the fact that there is no publicised case of such a trojan-ftp-hack at this moment. Do you realise that that will only happen AFTER people's machines are infected? Wouldn't it be nice if it DIDN'T happen to Mandriva? If I did, I wouldn't install those packages. Please don't imply paranoia on my part, I find it personally insulting and misplaced. If you believe it's not, read my message again. I'm not afraid _today_ that this will get people in trouble. I'm just explaining why _tomorrow_ it can. From the user point of view, sure, most likely. From the system point of view (how things are supposed to work), things have to change. There's no reason not to go for the proper methods if they are readily available. Well, if troubles started this would really be a black moment for Linux. It would be too late for those affected. And, worst of all, it could ALL have been avoided. By using the available mechanisms in the correct way. At which point today you won't have fun with your MDV06 system, considering all the packages you wouldn't be able to use. OOo2pre, to name a popular one. Again, based on what can a novice (who Mandriva really loves to have) make that decision? Remember, not everybody comes here and finds out what is what. Most people don't care, and if things break the system is at fault in their eyes. Which in this case can easily be defended. Oh and for sure those people (novices who feel they get burnt) either run back to where they came from or move to another distribution with the intention never to come back. I'm not telling anyone not to use packages with this issue, whereas I will advice anyone not to run as root. And I'm not paranoia even if you say/imply so. Please stop the ad hominem. Message on forum: hey I get his warning, so I didn't continue, now I don't have my software installed! Reply: Just ignore. Answer: ok, thanks, now it works. This is what I mean, and we have had those threads. If packages that Mandriva SHIP on the Powerpack dvd would not have this issue, no one would get conditioned to ignore warnings. This is incorrect. Read the manpage, search with google. Read my above comments. Signatures serve EXACTLY the purpose of offering a way to detect trojans and tampering etc outside of the doings of the packager. And I thought that Linux / Free Software was about the hightest standard... Anyway, it's just one single command to sign a package. Of course, if no one complains about it, this will be standard behaviour. So I'm doing the complaining, if no one else will. I do the translation for Mdv at my own expense. Firstly, I do get rewarded as do those packagers, with a VIP (~silver) membership. Secondly, if someone sends me a message that my translation is not up to standards, and especially how it can be done better, I will improve it. If people just decide not to use the localised version that I translate for, HOW CAN IT GET BETTER? I think Mandriva is getting very professional, the looks of the new site are great, they are doing better pr, and Mandriva Linux seems to be all that. I just think the i's should be dotted, and see no reason not to.

Well said John! I agree completely. Back on topic, when reading the start I wanted to comment immediately that the pc should be unplugged from the wall socket for some time - a few seconds should do. Remember, the power on some voltage lines (I thought 5V and 3.3V, but I'm not sure) is never down unless the machine gets switched off with the power switch at the back - if you have it, or by unplugging. There is nothing mysterious about that, just that some shadow caches and such don't get cleared properly. Note that if this were to happen the other way around - i.e. Linux being the majority platform, people would say: hey, MSWindows sucks, it can't start the network after I used Linux. Whereas in Linux, if this kind of thing happens/happened, they just fixed the drivercode to remedy the issue. So frankly, the MSWin driver should be fixed, don't complain here on a Linux board, complain to the manufacturer or MS for their half-baked system...

Cool! I put the note to the others, but not distrowatch - wanted to wait until my review is up before doing that. Thanks to Ladislav I guess.. :)

John, I agree that for now you can safely ignore. But - a time will come that you can't. Recall the events of the last 12 months: gnome server hacked into, gnu? server hacked into, and recently, a mozilla fansite spread trojans inside mozilla installers. The point being: Linux distributions are way ahead in terms of security. The signatures are intended to prove the source of the rpm. Now, think how one could get a trojan into a Linux system that's being used according to the rules. In Linuxland people don't just go to any website and download these cool programs, no, they just download from the official repositories. So setting up some website with lots of cool nice freeware programs - that just happen to be infested with trojans/spyware etc - is not an option. Infecting an ftp server that serves packages would be the way to go, because that's where people get their code that they have executed on their system. Replace an rpm on that ftp server may have people download this infected programs - but the signature will warn that it's not been packaged properly. So even this - currently non-existant - threat is already dealt with. If the sig is wrong, it means the source is not trusted. It gets better: even if the ftp server/mirror gets hijacked, and the keys on the server get replaced (ooh, lots of work) to match the keys of the infected packages, this would get noticed in two ways: first, all non infected packages that were signed with the correct key will now give an error, and chances are that the repository was added way before this particular piece of software gets installed at the users request. So the only ones who will get infected are those who add (update?) the repository, getting the new fake/wrong key and doing an installation of an infected package. All others will be in the clear. In comes Mandriva, training their users for years that one should just ignore the signature warning, and the whole idea is wide open to abuse. _THIS_ is why I don't like it. Really, in terms of security Linux is ahead of the curve, but if you take the shortcut you risk ending up in the water. Well, admittedly this is not a problem today, but the system is so well thought out that letting it break like this is really not good... Anyway, are you and/or others also having issues with packages from mdv06 main, contrib or the dvd?

I would think so - just to be sure, do a diff on your important files - then you'll know for sure.

John, I agree with the RW discs being problematic. I also agree with unbranded discs having higher risks. However, I disagree with your statement that it's misleading that +r is superior. Why? Because, technically, it is superior. (If I really need to get you the links that explain that tech stuff, let me know, I'll try to find it again, and it may take some time, ... but it is a fact.) Not that it matters, because to the computer user, it hardly makes a difference, mostly due to how people use dvd recordables and rewritables. Back to the -r statement that I would guess they are less good: another fact is that the dvd-forum, who ratifies the standards of dvd, including -video, -r and -rw, goes about validation and labeling differently from the +rw alliance. In effect, this means that it's relatively easy to get your products the label dvd-r(w). There is more and better quality control involved in getting a dvd+r(w) label on your discs. Taking unbranded discs is therefor a higher risk with dvd- than dvd+ whereas in the case of branded discs the risk to damage the brand is higher so the manufacturer/brand owner won't mess around with cheap bad stuff. In any case, what should not be left out in this picture is the immaturity of the dvd- and + standards - speeds are still being increased, new dyes are being used, etc. Basically, the recommendation for a type depends on the exact type of the media plus the burner. Some burners have better results with some discs, etc... So you may want to check out the results others have had with the same drive. To be honest, I don't trust dvds with my precious backups - there has not been any duration testing data of nonaccelerated tests simply because the media and technology are still so young. Accelerated testing has been done, high-uv and slightly increased temperatures in a controlled environment to imitate aging, with mixed results. The last such test that I have seen dates back a bit, and as new media are faster and use different dyes one can't draw many conclusions... Aside that, those tests are not related to + or - discs, but specific types (with specific dyes and write strategies)...

Quick question: do others using Mdv2006 also see so many Bad signature warnings? I get those a lot (really a lot), and not just on packages from remote ftp servers, also when I install things from the clubmember dvd. Am I the only one or is everyone already conditioned to ignore these? The latter would be really bad.. Please speak up!

Follow the info on my config page - you are a regular clubmember, therefor you deserve lesser service. Or so. Well, let's say that Mandriva had to come up with some distinction between the silver members and bronze/regular. So basically, you don't have those 5th and 6th cds. None of that should really worry you, the point is that you just have to set up urpmi as the page shows - if you look at the very top of this very page you're reading now, you will see a link called Easy-urpmi. Click it, and do the three step setup. In case of trouble open a topic on the respective subforum here. Enjoy Linux!

pmpatrick Ok, I have tried the recipe from that other thread to avoid high loads in kde by kded - they still happen, takes about 5 minutes, then it finishes whatever it's doing. I can still use the system, it's not too bad. Now I had the process mhonarc doing lots of stuff, actually the disk access is a few MB/s, cpu load is low, but the system becomes very unresponsive. Oh yeah: $ free total used free shared buffers cached Mem: 1035644 1021980 13664 0 1512 26440 -/+ buffers/cache: 994028 41616 Swap: 4898544 2473908 2424636 That's right: half my 4GB swap was in use. What???!! I have no clue who is converting mail to html, but I suspect kat - shut it off, no good. If this happens again, I'll urpme kat, just to see. beast2k, welcome to the other side of the 2005/2006 fence :D

pmpatrick, ok thanks for the info - I actually don't have that option of sharing folders - have to fiddle to get that working.. Will do and see if I have issues there too. Ix, I don't get any segfault, where are you fiddling? Look n feel - panels - appearance doesn't hurt my kde... Mirrors are getting updated, look at easyurpmi to see them being filled in for main and contrib... even updates are now available (some firefox stuff and then some, 21 files, 40 something MB in total...). If experienced people are holding the boat off, how are newbies going to be able to deal with things? Get with the program guys, 2006 is here so it's time to move on :D

Hi pmpatrick, NFS is too slow at boot, or if I put it differently, the boot is too fast! - so my network drives from the server don't get mounted - I put a mount -a in rc.local to still have them mounted before I log on (not that I need it as soon as I log on, I don't mount my homedir from the server or anything..). What's the issue for you? That hanging is really bad - not a showstopper per se, but darn, how does that stuff get by the alpha, beta, rc cycles??? Did you try boot options? BIOS options (pnp os no) ? Kernel mm? Do you have a clue as to how many machines are affected?

Thanks for the info. Side note: dvd+/-rw discs actually find their optimum after about 5x use and erase cycles. The problem is that they do get dirty which is much worse for a disc you still write to than one you just continue to read from...