joeclark

Hmm, I wondered if there was a trick to it. I don't know what happend with iptables, but doing my RPM search found openssl-devel and curl-devel for most other distros, so I figured that was probably a standard name. Guess I was wrong. Is there any way to discover names like "libcurl3" or is it just something you've gotta know? urpmi isn't exactly the most helpful at trying to find matches. Sorry if I sound annoyed, I just have these sorts of problems almost every time I want to do anything with RPMs.

So, I'm making up a nice fresh installation of Mandrake 10.1 to do some iptables development work. I needed to add the following packages: iptables-devel openssl-devel curl-devel urpmi did not find any of these (the machine is offline so it would have only searched the CD databases). I looked at rpmfind.net and found the right version of iptables-devel. However, I could not find anything at all for Mandrake/Mandriva for openssl-devel and curl-devel. So, I'm building from source. So much for package management. This is a good share of the reason that people like me think RPMs are evil -- the distribution just always seems to be missing something I need to have (or maybe gives it some non-guessable name or something). Any ideas why 2/3 development packages don't seem to exist?

One more report: Well, I found a workaround. My code was setting the LookAndFeel to the SystemLookAndFeel. By *not* trying to set it, it just uses the Swing default, which looks good enough to me and is plenty better than the "ugly" style it was using previously. So..that solution works for me, although there's still a question about why the themes don't work correctly in Gnome/GTK.

Update: Okay, Java sees the theme change after logging out and back in (or restarting X, etc). Question: What themes *does* Java support?

When I start a Java GUI application with JDK1.5 under Gnome on Mandrake 10, I get the following theme error from the JVM: /usr/share/themes/Galaxy/gtk-2.0/gtkrc:43: Engine "galaxy" is unsupported, ignoring. The resulting GUI is then drawn in some basic (fairly ugly) fashion. Can someone explain to me simply how to change the theme (the theme manager in Gnome didn't help) so that Java does not complain about it? I don't really want to upgrade gtk2 or downgrade Java -- there should be a simple way to fix this I would think, but I'm not sure what it is. Thanks.

I have a Mandrake 10.0 machine running as an NFS server. I am accessing this server from Red Hat 7 clients. It seems that whenever I try to do anything with an NFS mount, the client machines hang. I tried the sync/async server option, but both act the same. Looking with tcpdump, it seems that the server simply stops responding to requests, although there is no syslog entry reporting a problem on the server. Any ideas? - Joe

Hi there. I frequently use CDSpeed on Windows to verify that burned discs are sector-by-sector readable. I would like to find something with the same functionality on Linux, but so far I haven't come up with anything. My requirements are: * I don't have the original data available on Linux (it was burned on Windows) * I don't have an ISO image or an md5 sum (I burned it direct-to-CD) Anyone know of an easy way to accomplish this with the given requirements? For "flame bait", remember this is extremely easy to do on Windows (with CDSpeed). :-) - Joe

Really? 10 Official has this same problem? Are you sure? I thought it was fixed (where fixed == worked around the Windows issues or whatever). I remember "using Windows free space" to dual-boot install 10 Official on a machine and was pleased to find that it didn't do evil to it...maybe I got lucky or something?

I have Mandrake 9.1 installed on a server. I would like to be able to disable the automatic logout "feature" of sshd (or msec) that apparently is the default...it logs me out after about 10 minutes of inactivity. I think there's an easy way to un-configure this, but in my web search I couldn't find the answer. Can anyone here help? Thanks. - Joe

I am running Mandrake 10 Official... I have a Java program that, well, may or may not have a memory leak in it. I'm a little overwhelmed with all the memory-related numbers: Java used and free numbers and physical vs virtual and virtual vs swap (same thing right?) and Java vs Linux numbers and all. But in any case we are running a Java program that shows a fairly steadily increasing "RSS" size in ps, at least until it consumes about 170 MB of RAM. However by doing a -verbosegc option to the VM it doesn't obviously appear to be actually using more RAM in Java. In any case, at some point Linux/Java finally pages some of the Java memory out to disk or something, and the RSS size goes down to 125-150 MB. I believe it is after this that I frequently see the X process consume upwards of 80-90% of the system RAM over a period of 5 or so minutes. Then, when all free memory (as reported by cat /proc/meminfo, "MemFree" line) is exausted, and the swap starts to be allocated, X dumps a ton of RAM in less than 10 seconds and goes back to appx 150-200 MB. In a current running configuration, this periodic effect has occurred about four times now, and each time the "SwapFree" has gone down by at least several megabytes. So, I'm guessing that in the long term some memory is getting leaked, and eventually the system RAM (virtual + physical) will be exhausted. Note that actually when this huge memory dump occurs, the Java process memory also decreases. So far it seems to have decreased each time X did this strange thing, and Java's usage is now down to 100 MB or so. I can provide some more specific data on this if anyone thinks it would help. Anyone have a clue what might make X act so strangely?

Thanks for your input...with the old calls in place, it seems much happier. No problems after running under normal operating conditions for 2-3 days. I'm going to put a new thread up here though about X & Java memory usage....at some times XWindows get into a funny state and sucks up just about all the system RAM...I'll describe in the other thread.

Well, we're going to remove the calls that seemed to cause the problem (Graphics2D.draw(Ellipse2D)), and replace them with the older fillOval method (we had previously used fillOval and didn't see a problem). I am hopeful that this will "fix" the problem (or be a sufficient workaround). I would consider using another VM, but hopefully this fix will work. One other interesting thing related to MDK10: While this Java application is running, sometimes I see the X process using up to 90% of the system RAM. On a machine with 1 GB of RAM, this is surprising to me, esp when it seems the nominal usage of RAM for X is about 2-3% of the total. Any ideas on if this is some kind of X bug and if it might be related to Java somehow, or if it's another issue entirely? Thanks your your help on this topic. It's been good to "vent", even if there isn't a clear solution.

I installed the newest Java SDK available (1.4.2_04), and it has the same problem. What's left but to install a different OS? This is really going to annoy me that Mandrake 10 official has some Java bug...is there any way to even know what OS/library component the bug is in? Is it really a MDK bug, or a kernel bug, or an X bug, or ..... ? I guess maybe we'll go back to (old) 8.2, just because we know it works. About memory, again: When it crashed this time, Java VM was using about 94 MB of RAM. Yes, that's a lot, but not unreasonable in our situation. It may have something to do with memory, but the fact is it seems to work fine on machines with less memory as long as they're not running Mandrake 10. Grr. Another one of those "stupid linux" days. (Don't worry, I have plenty of "stupid Windows" days too.)

It's from a self-extracting bin...probably the same bin file used to install on Mandrake 8.2 We've been dealing with some memory issues as well, but the Mandrake 10 machine should have plenty of free memory to spare.... In any case, it's the only machine we've seen this problem on, and it probably has the most memory of any machine we would try it on.

We are using Sun's JDK for Linux, version 1.4.2-b28, on Mandrake 10 Official. Our Java application runs basically all the time, and we interact with it frequently. It is a graphical application utilizing a simple mapping architecture. Our problem is that about once every day or two, the Java virtual machine crashes with a seg fault outside of the VM code. This is obviously bad, but I can't find a similar crash report by searching Google. Perhaps this should be on a Java forum, not a Mandrake forum, but I thought I'd start here. Also we run the same Java application with the same exact version of Java on a Mandrake 8.2 machine, and have never seen this error, so I'm thinking it might point to something in Mandrake. Would you suggest trying a newer version of Java? Or a Java code change? Or (sigh) a different version of Linux? In the stack trace (printed below), the call originates from a Graphics2D.draw(Elllipse2D), if that helps. Here is part of the error that the VM dumps out when this problem occurs: An unexpected exception has been detected in native code outside the VM. Unexpected Signal : 11 occurred at PC=0x4DA9C1E4 Function=(null)+0x4DA9C1E4 Library=/usr/java/j2sdk1.4.2/jre/lib/i386/libdcpr.so NOTE: We are unable to locate the function name symbol for the error just occurred. Please refer to release documentation for possible reason and solutions. Current Java thread: at sun.dc.pr.PathStroker.endPath(Native Method) at sun.java2d.pipe.DuctusRenderer.feedConsumer(DuctusRenderer.java:232) at sun.java2d.pipe.LoopPipe.getStrokeSpans(LoopPipe.java:259) at sun.awt.X11Renderer.draw(X11Renderer.java:243) at sun.java2d.pipe.ValidatePipe.draw(ValidatePipe.java:114) at sun.java2d.SunGraphics2D.draw(SunGraphics2D.java:2128) *** Moved by Phunni

The saga... So I got the Official ISOs overnight from Mandrakeclub/BitTorrent. I see that they include drakxtools-10-29. I tried doing the diskdrake --change-geometry=hda=255,63 command, and it does something for a bit, and returns to command prompt, but upon rebooting the problem is the same. This was expected, since I'd already tried drakxtools-10-29 as a cooker release on CE. So I am close to the end of my patience with this. I suppose I'll just reformat the thing soon. I agree that it just plain seems "screwed", although I don't get why the supposed fixes don't fix it -- that's the nature of computers I guess. :-) It has become more a philosophical endeavor for me than anything else: "Can Linux repair what it broke". In this case, or due to some random thing I did after the fact, I guess the answer is no. This does not make me happy, but I put up with it. It's not completely unexpected, I know, especially with a "prerelease" version, but let's just say that things like this don't help the reputation of Linux. PS: I did a dual-boot thing with a pre-existing Windows installation at work today with the Official CDs, and it *didn't* break, so I'm guessing they fixed the root cause bug (as I expected). PPS: One last effort: I noticed tonight that when Windows is blinking cursors at me, it seems to be accessing the floppy drive -- not actively really, but the light is on. Any clues based on that?

Anybody know yet whether the Mandrake 10 "Official" version *fixes* this problem (ie if I do an upgrade, will it fix my partition table)? Hopefully it will....it's really "special" that this boot issue is not even mentioned on the errata page for 10CE (at least it wasn't last I noticed).

Yeah, the reason here, though, is that something is screwed up in the MBR/partition table due to a Mandrake bug..I don't think that it has anything to do with the actual Windows files on the Windows partition. When I added the Mandrake partitions, it took space at the end of the physical disk -- the Windows partition is still in the same location where it used to boot from.

**Update** I think I figured out some things... in my ifcfg-eth1 file, I had parameters like WIRELESS_PARAM. These are used in ifup to set the iwconfig properties. However, ifup-wireless uses different values of just PARAM without the wireless. I think that ifup-wireless is careful to set the essid last, whereas ifup is not. So..I think the magic is in using the right set of variables....by the way...why are there two sets? If ifup-wireless sets stuff, why does ifup also set stuff? Oh, and why is "NICKNAME" set to hostname? Wouldn't it be better to allow the user to set that in ifcfg-eth1 too? Another thing: it seems that at the very end of ifup-wireless I have to add at the end a sleep for at least 1 second, followed by an ifconfig down/up sequence. Only then will the card initialize itself properly so that I can ping the router. Any ideas why that is? There was a reference online somewhere that said this was necessary (and I saw that it wouldn't work without it). Finally, on startup, the "[OK]" list shows that eth1 failed to load, but when I finally get to a console that i can try it, it seems that it DOES work now. Also, if I run network restart manually, it takes appx 5 seconds to run (with the delays I guess), but at bootup it takes almost no time. Is the startup sequence not waiting for a valid response from the network script before it goes on and so assumes it failed? Note: Wouldn't it be easier if there were an ifcfg-sample that included all the options a typical user might want to add in. One extra file might make a lot of people not waste so much time. And while I'm at it, having sample commented-out lines in /etc/exports and /etc/sysconfig/static-routes (and probably lots of other places) would be nice too. All three of these files have now annoyed me because they don't provide samples.

/etc/pcmcia/config -> this file is way long (2336 lines)...I've attached it below. /etc/sysconfig/pcmcia: PCIC=yenta_socket PCMCIA=yes /etc/modprobe.conf (I'm using kernel 2.6.3): alias eth0 via-rhine alias sound-slot-0 snd-via82xx options acx100_pci use_eth_name=1 firmware_dir=/lib/modules/acx100_fmwe install scsi_hostadapter /sbin/modprobe imm; /sbin/modprobe ppa; /bin/true install snd-via82xx /sbin/modprobe --first-time --ignore-install snd-via82xx && { /sbin/modprobe snd-pcm-oss; /bin/true; } install usb-interface /sbin/modprobe usb-uhci; /bin/true remove snd-via82xx { /sbin/modprobe -r snd-pcm-oss; } ; /sbin/modprobe -r --first-time --ignore-remove snd-via82xx I'm not sure whether the firmware option was necessary or not...I did a couple things to the configuration to get it to work to start with, but I was un-smart and didn't write down the exact steps and changes. Thanks for your help. config

About installing Windows first: I agree. I like lilo, and would prefer it to load either Windows or Linux. That has worked fine in the past, but with the crummy buggy partition program in MDK10 stuff gets broke. My saga: Always something....I managed to get the updated drakxtools (v 10-29), and ran the command described in a bugzilla report on this problem (http://qa.mandrakesoft.com/show_bug.cgi?id=7959). That didn't help. Then I tried the fdisk /mbr from FreeDOS, and it did a good job of deleting lilo, but it didn't help make Windows boot at all. I then tried to boot the Windows recovery console, but the first thing it did (after taking forever to load a bunch of driver files) was ask for the Administrator password -- if I remember right, the only users in XP Home were myself and the "Owner" account. Great..so that didn't work. Then I tried using Mandrake to restore the Windows bootloader (after reinstalling lilo). It installed it I guess, but Windows still wouldn't boot. So, I finally quit and just put lilo back on...at least lilo will boot Mandrake still. So, the only thing I haven't tried is the "LBA" switch. However, as I mentioned before, my laptop has one of those software (hard disk) BIOS thingies, so the real setup in the bios has very few options. I tried to run the diagnostic tool to run advanced setup (or whatever they call the hd-based one), but it seemed to be inaccessible (but I KNOW the hidden partition is still there). Whatever, computers are stupid. One more thought: would it do any good to try and reinstall the MBR using **Mandrake 9.2** rescue mode? Would that fix it, or do nothing, or make it worse? Is the problem even the MBR, or is it in the partition table itself?

I have a DWL650+ wireless card, controlled by driver acx100_pci, with Mandrake 10. When the system boots, the interface (eth1) is configured with wireless parameters and IP address, but I am never able to ping the wireless router initially. After startup completes and I login as root, I've tried the following: /etc/init.d/network restart and ifdown eth1; ifup eth1 Neither of these "connects" the card reliably. The only reliable way to make the card work (has worked every time) is to do the following: ifconfig eth1 down; ifconfig eth1 up; ifconfig eth1 ipaddr From then on, the card works fine. I would like to figure out WHAT ifconfig down/up does that even ifup/ifdown do not do, and why the card does not initialize appropriately on startup. Any clues? PS: If Mandrake hadn't messed up my partition table (making Windows unbootable), maybe I wouldn't mind this bug, but if Linux is going to be my only available installation for awhile, I'd like to have it work *right*.

The link you mentioned lists 3 RPMs, but they are not clickable (they give ftp errors). What's wrong the Mandrake mirrors tonight anyway? Are they all broke or something? Um, I think the error the original post-er was mentioning was that they have the files, they just won't install because of some weird dependency. I'm having the same problem, but I've added too many similar posts on too many threads here tonight, so I'm giving up and calling it a night. BadMoodFactor = TimeWasted * Hours Past My Bedtime

I have a cooker source set up (one of the very few cooker mirrors that actually seems to work), but it lists a similar error when trying to install the file. BTW, I trued the site you suggested, but it also produced an invalid ftp path (ftp.orst.edu/.1/mandrake-devel : path does not exist, I checked with anonymous ftp). Bleh.

on MDK10CE, I am trying to update drakxtools to a cooker version (10-29), and to do so I need to install perl-Libconf. I found a package for it, but when I try to install it, I get these errors: [root@localhost root]# rpm -i perl-Libconf-0.29-1mdk.noarch.rpm error: Failed dependencies: perl(Libconf::Glueconf::Hosts) is needed by perl-Libconf-0.29-1mdk perl(Libconf::Glueconf::Network) is needed by perl-Libconf-0.29-1mdk perl(Libconf::Glueconf::Passwd) is needed by perl-Libconf-0.29-1mdk perl(Libconf::Glueconf::Postfix) is needed by perl-Libconf-0.29-1mdk perl(Libconf::Glueconf::Resolv) is needed by perl-Libconf-0.29-1mdk perl(Libconf::Glueconf::Shell) is needed by perl-Libconf-0.29-1mdk [root@localhost root]# rpm -q perl-Libconf package perl-Libconf is not installed Now, it sure seems to me like it's trying to tell me that I need some part of Libconf installed before installing Libconf (!!!) What do these errors actually mean, and how can I fix them? Thanks. - Joe