Re: entropy
Luca wrote: Hi all, if I simply write to /dev/random, will that increase the entropy of my system? (I'm assuming that the data I'm writing are random and that somehow I got them). Wikipedia says so. My tests say no. In particular this brutal approach does not increase the entropy cat /dev/urandom /dev/random (it is stupid to do that, I know, but it's just a test) You could investigate on how rngd works. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: entropy
Tony Nelson wrote: On 10-01-07 12:40:02, Roberto Ragusa wrote: Luca wrote: Hi all, if I simply write to /dev/random, will that increase the entropy of my system? (I'm assuming that the data I'm writing are random and that somehow I got them). Wikipedia says so. My tests say no. In particular this brutal approach does not increase the entropy cat /dev/urandom /dev/random (it is stupid to do that, I know, but it's just a test) ... `man 4 random` says that the current entropy can be read and written from /dev/urandom, not /dev/random. This is used to preserver entropy across reboots. That's true. But as far as I can see neither writing to random nor to urandom will increase the entropy availability. After checking the sources of rngd, I found it uses a specific ioctl: ioctl(random_fd, RNDADDENTROPY, entropy); So I think Luca can inject entropy by using the same ioctl in his own application, or by using rngd in some way (you can tell it where to take entropy from). -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Tar oddity...
DB wrote: On 12/22/2009 09:51 PM, Roberto Ragusa wrote: -variable numbers of md5sum errors running on the F12 laptop (for example, will run 3 times with same checksum then 1 error); variable cmp errors(again, one or two clear runs then a pile of miscompares, then again several ok runs); That is really bad. If the machine is unreliable, any kind of hard computation (such as data compression) will have problems. Maybe the problem is caused by something bad in the F12 distro (such as new graphics drivers). I suppose you tried the LiveCD on the laptop, and it looks like F11=good, F12=bad. If /usr/bin/ark is fubared, any idea how I can best put it right? yum reinstall of the correct rpm (kdeutils?). Try also rpm -V to check. But if md5sum and cmp randomly fail, the problem is not ark. The machine is unstable, for hardware or software reasons. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Recover Root Password on FC 11 and Missing GRUB Screen
Hosea Phiri wrote: My surprise, the machines boots differently. I noticed one major thing that looked different from other versions of Fedora I have used before. It does not bring up the Grub menu. It does not even show the services startup. It goes straight into login prompt bypassing all other stages which I guess run from background. Any I dea? Keeping a key pressed at the GRUB execution instant. Or, booting from a rescue CD/DVD. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: installing 64-bit kernel on a 32-bit system (nouveau issue?)
(warning: added cross posting to fedora-devel) Ramesh.R wrote: You can use 32 bit OS in 64 bit processor. 32 bit address bus will use 64 bit. MSB 32 bits will be idle.. But for the case, 64 bit OS in a 32 bit processor is not possible by theory. No one is talking about that. You are not the only one in this thread to have misunderstood. Suppose you have a 64 bit processor. You can run: (a) 64 bit kernel + 64 bit apps: that would be a pure 64 bit system (b) 64 bit kernel + 64 bit and 32 bit apps: that would be a multilib system, where you keep some 32 apps for some reasons A normal Fedora installation will give you case (a) or (b). Now, consider this: (c) 64 bit kernel + 32 bit apps: this is simply an extreme case of (b), a 64/32 system where every app is 32. Case (c) is interesting because: - you can switch a 32 bit install to this mode by simply installing a 64 bit kernel (and switch back at grub level any time you want) - the 64 bit kernel can handle all your memory better (faster) than 32 or 32+PAE kernel - you avoid the increased memory consumption of 64 bit apps (pointers are wider; there is big debate how much this impacts performance and if it is able to demolish the other improvements of x86_64 such as more regusters and SSE2 guaranteed avalability). Add to this that when you run 32 and 64 bit apps together you have both versions of the system libraries in memory, so the mem usage is higher. Finally, the discussion is: case (c) _SHOULD_ work perfectly in theory (see case (b)), but apparently there are a couple of bad spots for things no one ever run in 32 bit mode on 64 bit kernel. The first one I heard is the Nvidia closed source driver (ok, we already know closed source = unfixable problems), but this thread seems to suggest that nouveau has a similar issue. My suspicion is that it is just an untested area and a fix could be done easily. In any case, it looks like a show stopper for nvidia users. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: installing 64-bit kernel on a 32-bit system (nouveau issue?)
(warning: added cross posting to fedora-devel) Ramesh.R wrote: You can use 32 bit OS in 64 bit processor. 32 bit address bus will use 64 bit. MSB 32 bits will be idle.. But for the case, 64 bit OS in a 32 bit processor is not possible by theory. No one is talking about that. You are not the only one in this thread to have misunderstood. Suppose you have a 64 bit processor. You can run: (a) 64 bit kernel + 64 bit apps: that would be a pure 64 bit system (b) 64 bit kernel + 64 bit and 32 bit apps: that would be a multilib system, where you keep some 32 apps for some reasons A normal Fedora installation will give you case (a) or (b). Now, consider this: (c) 64 bit kernel + 32 bit apps: this is simply an extreme case of (b), a 64/32 system where every app is 32. Case (c) is interesting because: - you can switch a 32 bit install to this mode by simply installing a 64 bit kernel (and switch back at grub level any time you want) - the 64 bit kernel can handle all your memory better (faster) than 32 or 32+PAE kernel - you avoid the increased memory consumption of 64 bit apps (pointers are wider; there is big debate how much this impacts performance and if it is able to demolish the other improvements of x86_64 such as more regusters and SSE2 guaranteed avalability). Add to this that when you run 32 and 64 bit apps together you have both versions of the system libraries in memory, so the mem usage is higher. Finally, the discussion is: case (c) _SHOULD_ work perfectly in theory (see case (b)), but apparently there are a couple of bad spots for things no one ever run in 32 bit mode on 64 bit kernel. The first one I heard is the Nvidia closed source driver (ok, we already know closed source = unfixable problems), but this thread seems to suggest that nouveau has a similar issue. My suspicion is that it is just an untested area and a fix could be done easily. In any case, it looks like a show stopper for nvidia users. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: reboot cycle, no log messages
Joel Rees wrote: yum provides libfreebl3.so tells me that nss-softokn-freebl is installed. Trying an erase to re-install it tells me that most of the OS and most of the apps seem to be dependent on it, one way or another. Hmm. Bypass yum. rpm -V nss-softokn-freebl or/and rpm -Uhv --replacepkgs --oldpackage nss-softokn-freebl.rpm Are you on 32 or 64 bit? If 64, maybe the libs are incomplete. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: installing 64-bit kernel on a 32-bit system
john wendel wrote: Just for fun, on F11 32-bit system (not tried on F12), I downloaded the latest F11 64-bit kernel package and installed it with rpm --nodeps --ignorearch --force kernel package name It installed OK, since the kernel is pretty isolated from the rest of the system software. Booted into runlevel 3, and it worked fine. Now, X won't start because it needs the 64-bit nouveau driver. I don't know, but I suspect that you will need to install the entire 64-bit X server package(s). This will force you to install at least the 64-bit libc package. Interesting. A similar problem had been described for the Nvidia closed driver; same thing with noveau? not a good thing: the kernel-X interface is expected to be cleaner in the open source driver... -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Hard Drive upgrade with Fedora 12 installed
William M. Quarles wrote: The XP and the PC DOS I can handle, but I am unsure of how to transition the Fedora 12 installation due to the fact that I know next to nothing about SELinux, how it affects /etc/fstab since Fedora started enabling SELinux by default years ago, nor how to copy/rebuild the /dev and /proc trees on the new hard drive. Any help you can give me would seriously be appreciated. SELinux: just disable it before the operation, get your system working on the new drive and then think about reenabling it. /proc: there is nothing to rebuild, the files are all virtual. /dev: nothing special to rebuild, all automatic Your real issues are: - when you copy the data (better if using a rescue disk) to the new partitions you will have different partition name and maybe different filesystem UUID; this will affect /etc/fstab and grub.conf - grub has to be installed on the new drive and its config modified - the booting process is also dependent on your initrd in /boot Did you search for some detailed tutorial on how to do this? I guess this process has been described many times in the past. It is not a difficult operation: you just copy all the files and then repair the boot process. As always, the devil is in the details. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Pondus license change GPLv3+ - MIT
Ralf Ertzinger wrote: Hi. On Sun, 27 Dec 2009 17:43:05 +, Ikem Krueger wrote That's doable? o.O The copyright holder can relicense the code however they see fit. What they cannot do is retroactively remove the GPL license from old versions. Relicensing is complicated when there are a lot of authors (copyright holders), because it is necessary to get the approval of everyone of them, with complications for unreachable people and dead people; in the past discussions on the GPLv2-GPLv3 relicensing for the kernel, there was an additional opinion that this requirement is not as stringent as it appears. In any case, relicensing a small project with a few authors (or just one) is as easy as editing a couple of files (source, webpage, ...). -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: md raid 5 not working
Kanwar Ranbir Sandhu wrote: I realized I had once used the entire drives in a md RAID 5 set instead of building the RAID 5 on partitions. I had outdated md superblocks on /dev/sd[bde]! In fact, I forgot to mention in my reply that this was a little suspicious: Preferred Minor : 2 I guess you had md2 on entire disks, and then switched to md3 for partitions. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: md raid 5 not working
Kanwar Ranbir Sandhu wrote: /dev/sdb: 0 0 8 160 active sync /dev/sdb 1 1 001 faulty removed 2 2 8 642 active sync /dev/sde Events : 13144 /dev/sdd: 0 0 8 160 active sync /dev/sdb 1 1 8 481 active sync /dev/sdd 2 2 8 642 active sync /dev/sde Events : 11958 /dev/sde: 0 0 000 removed 1 1 001 faulty removed 2 2 8 642 active sync /dev/sde Events : 13146 Your three disks have three different ideas about the state of the RAID array. The events counters are probably saying that: - sdd was removed from the array at 11958 (faulty!) - sdb was removed from the array at 13144 - we are now at 13146 I think you should try to assemble the array in degraded mode with only sdb and sde, which have almost the same age; sdd is probably stale and should be readded in a second moment (assuming the disk is really working). I'm not sure about what options are needed to assemble in a degraded mode by choosing only sdb and sde, so I will avoid giving you a bad advice which could destroy your data. :-) You may retry the rescue CD; is the array started in degraded mode? If so, you can try to readd the missing disk and let it sync. Maybe everything is clean on next boot. But be careful. As soon as an array is reassembled, run a read-only fsck to be sure you are reassembling something consistent (I personally would not trust any array involving sdd). -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: IBM Netfinity 5000
Ashley M. Kirchner wrote: I have an old Netfinity 5000 server that used to run WinNT and I'd like to install Fedora on it, any recent version would do (basically, from 10 up). I started with a Fedora 12 DVD and it booted partway ... froze, then the screen said 'Cannot find root partition.' So I decided to try a Fedora 10 disc instead, same result. Tried using an F10 LIVE KDE disk, same thing. It just won't boot up. Hardware wise, it has 2 SCSI drives in it, and an IDE DVD drive. A dual head Matrox card has been put in to replace the original planar video on the board. Any suggestions where I should start diagnosing this beast? Thanks! It sounds like a problem in detecting the DVD drive. Strange, if it is just IDE. I woud try some other Live distribution, just to have more data (does if fail? what kind of controller does it see? ...). Old glorious Knoppyx, for example. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Tar oddity...
DB wrote: Just tried to copy the file on the F12 laptop... only 324MB of 1.7GB get copied. I'm beginning to wonder whether my F12 installation is without error, as I keep losing my (KDE) panels. Each time I add a new one (so that I can open these files etc) the widget-adder show x copies of the widgets being in use - but I can't find a way of materialising the invisible panels Also, if I zoom out I see 2 screens, one plain blue, one the (for want of a better term) blue blobby one - or is this somethinf=g to do with disabling the laptop screen with Marko's xrandr script??? Confusion reigns supreme! It looks like you could have some stability issues. I would try some basic test for your hardware. For example: while true; do md5sum somefilewithdimensionabout50MB; done Are the checksum all the same or do they change? If they change, causes could be: overclocking, bad power supply, bad RAM, bad CPU or software issues (bad kernel, unstable drivers). Try then with a bigger file, so it will have to read the disk. You can try the external disk, so you will test the USB part too. Create a file of 500MB with: dd if=/dev/urandom of=myfile bs=1048576 count=500 (replace 500 with what you want), then checksum it while true; do md5sum myfile; done use a dimension bigger than the RAM you have on your machine, so you defeat the caching that Linux will try to do. Another test could be [create some file called a] cp filea fileb while true; do cmp -l a b;done The idea is do have the machine do a lot of work for which you already know the results. If you get a bad answer (a and b are different??) you know the machine is not working well. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Routing with 2 ISP
David Hláčik wrote: And the finally my questions are : 1) Is there a good tutorial / howto for using iproute on the internet, except of the LARTC.org 2) Can i utilize by tools of Fedora, to have my configuration (with second routing table, using ip ) somehow stored - to be permanent when I will do machine restart? I mean there are networking-scripts /etc/sysconfing/network-scripts which can handle, IP assigment, virtual LANS, aliases even static routes. Can they handle advanced routing as well? Thank you milion times. I have a setup similar to yours. In my case I also have a script checking the two providers and switching the connections on the other one when one is not working. You can find documentation around on iproute, I think it was the lartc.org site I used for reference. As the integration with Fedora network scripts I think there is no way to have such a complex configuration in a clean system. This is on Fedora 9, but I doubt things are improved in this field in newer version, considering the rampant NetworkManager-style approach to networks. In my case I have the ifcfg scripts doing some basic configuration which will not enable significant traffic (for example forwarding=0, no routing, no default gw). Then I have a custom script called by rc.local which does everything: iptables firewalling, iptables MARKing based on --match conntrack, ip route routing based on fwmark. I tried to achieve external visibility on both my public IP addresses for incoming connections, but something is still not working; it looks like there is a problem with the tracking of the incoming connection because it is not established yet. I stopped spending time in fixing this, as my target was to have the fail-over part working. My config is actually more complex than what I've described (vmnet networks, port redirections, traffic shaping, traffic stats,...). If we are lucky, someone really expert on this will reply in this thread. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Tar oddity...
DB wrote: The reason I went with tar tvh was (to try) to check the contents of the file after open with ark in Dollphin spat out the errors. I guess that actually trying to extract the files when the table of contents fails would not be any more successful? Run md5sum F11_Home_Dave_20091217.tar.gz on both machines. If you get two different results, something bad is happening. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: What's that utility to monitor disk i/o again?
Marcel Rieux wrote: mount a /sys/kernel/debug -t debugfs [...] All references at Google's, only 9 of them, say mount a. How come it's not mount -a a is the device containing the filesystem. For this kind of filesystem it is unused so you can put everything there (the same happens for tmpfs, proc, sysfs, devpts). 1. Manually mount after each boot: % mount -t debugfs debugfs /sys/kernel/debug In this case the second debugfs could have been a or whatever else. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: physical RAM restriction in Fedora 12 (32 bit and 64 bit)
Hiisi wrote: 2009/12/18 Roberto Ragusa m...@robertoragusa.it: Kevin J. Cummings wrote: On 12/17/2009 12:51 PM, Mark Ryden wrote: --SNIP-- And the PAE kernel will be installed by default. So the 32 bit Fedora could be considered without limit too. By the way, will I see any speed burst if I replace PAE kernel with the ordinary one (not PAE) on 32-bit F-11? My computer has only 2 GB of RAM. The non-PAE version used to be preferred. But experts say the performance impact with PAE is minor (at the almost unmeasurable level) with recent versions. There is also some small security advantage with PAE (NX protection, IIRC). -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: disk moves from /dev/sdd to /dev/sde
Wolfgang S. Rupprecht wrote: Thanks. I (at the /etc/fstab level) do use a label that goes with the drive, mainly the lvm VG and LV names. The problem is that the lvm mapper *internally* must still use the name that the drive had when the computer first booted. It gets confused when the drive gets renamed at runtime (long after boot). Ditto. I have my machine dual ported and I'd really like eth0 to be the internal ethernet and eth1 to be the external. (Some programs like multicast progs seem to default to the first ethernet and it would be nice for me not to spew packets towards the public internet.) I think you can address both problems with udev or HAL. I've done it in the past, but these things are changing too fast and with almost no documentation around. It could be that on newer Fedora things are different. My /etc/udev/rules.d/70-persistent-net.rules contains (xx where I removed my data): (on Fedora 10) # PCI device 0x8086:0x1049 (e1000) SUBSYSTEM==net, ACTION==add, DRIVERS==?*, ATTR{address}==00:xx:xx:xx:xx:xx, ATTR{type}==1, NAME=eth0 # PCI device 0x8086:0x4230 (iwl4965) SUBSYSTEM==net, ACTION==add, DRIVERS==?*, ATTR{address}==00:xx:xx:xx:xx:xx, ATTR{type}==1, NAME=wlan0 # PCMCIA device 0x0202:0x021b (axnet_cs) SUBSYSTEM==net, ACTION==add, DRIVERS==?*, ATTR{address}==00:xx:xx:xx:xx:xx, ATTR{type}==1, NAME=eth1 I created a /etc/udev/rules.d/11-ROB.rules for the external usb disks containing: (on Fedora 5) BUS==usb, SYSFS{serial}==5743, SYSFS{product}==External HDD,KERNEL==sd?1, NAME=external_bup_disk BUS==usb, SYSFS{serial}==5744xx, SYSFS{product}==My Book ,KERNEL==sd?1, NAME=external_bup_disk2 The serial has to be found in the lshal output (a lot of output, not easy). In both cases the NAME part is where you decide what you want to have in /dev. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: physical RAM restriction in Fedora 12 (32 bit and 64 bit)
Kevin J. Cummings wrote: On 12/17/2009 12:51 PM, Mark Ryden wrote: My question is: 1) In Fedora 12 32 bit default installation , does the kernel knows more than 3 GB of RAM ? what is the limit ? The same. It uses what the BIOS tell it is available, unless you run a PAE kernel. The Physical Address Extensions allow you to address the full amount of your RAM. And the PAE kernel will be installed by default. So the 32 bit Fedora could be considered without limit too. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: su hangs for 30 seconds
Joachim Backes wrote: On 12/05/2009 01:32 PM, Hiisi wrote: 2009/12/5 Wolfgang S. Rupprechtwolfgang.ruppre...@gmail.com: As of a day or so ago su has started hanging for 30 seconds. So has the lock screen. Jiggling the mouse unblanks the monitor and shows me the backdrop picture but the password entry box doesn't appear for 30 seconds. I don't believe I mucked with anything PAM related, but there were a few yum updates in the last few days. Is anyone else seeing this? -wolfgang -- Wolfgang S. Rupprecht If the airwaves belong to the public why does the public only get 3 non-overlapping WIFI channels? I have the same problem for a couple of month (don't remember exactly how long it is) on my F11 (32 bit). I've asked it already on this list but had no response. I had similar problems in the past (with sudo / not su), and the reason was an error in the network controls (I tried to change the hostname by editing /etc/sysconfig/network, but forgot all other places to edit). This kind of delays are often DNS timeouts. If the network configuration is wrong, trivial things like printing last unsuccessfull login on 02-12-2009 from abcd.example.com take 15-30-60 seconds. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: anyone noticed this odd firefox glitch?
Robert P. J. Day wrote: i've seen some trivial but fairly new oddities as well. as i mentioned before, the scroll bar doesn't seem to act consistently. once upon a time, if i clicked way down the scrollbar to page down, firefox would, well, page down. once. now, fairly regularly, it will blow through a massive amount of scrolling down. immediately thereafter, though, it will go back to what i recall as normal behaviour. Are you on x86_64? I've heard your issue mentioned somewhere as a 64 bit problem. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Fedora 12 x86 DVD images
Sir Gallantmon wrote: Why not label it x86_32 instead of i386? That is far less confusing and illustrates that it is 32-bit on the x86 architecture, since x86_64 says it is 64-bit on x86 architecture. Because x86_32 is not an architecture name. You are just creating it from x86_64. 32 bit is i386 or IA32. 64 bit is x86_64 or AMD64 (BTW, I would have preferred AMD64 to be more used for 64 bit, as AMD should be given credit for the creation of the architecture, in contrast to Intel which gave us the disaster called IA64). -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: Upgrading from FC8 to F12 - please help
Andrew Junev wrote: I apologise for such a messy description. If there's anything I can provide or clarify - let me know how I could do it. I would really appreciate any hints on how to get my system back... Update: now while trying to boot to an old FC8, I get an error: /sbin/init: error while loading shared libraries: /lib/libsepol.so.1: invalid ELF header Kernel panic - not syncing: Attempted to kill init! Doesn't look good... :-( I would prefer to continue with an upgrade to F12 if possible, but I don't know how to do it and I'm not sure it is possible at all... I think the fact that sda and sdb are now 1TB indicates that the names got changed (extra controller vs motherboard). I'm not sure about the sda sdb RAID1 configuration. Is it possible that sometimes you use the couple as a RAID (old Fedora) and sometimes you use one of them. If the partial upgrade happened on only one of them the mirror has been broken and all kinds of problems can happen. I would suggest you to separate your recovery in steps. First thing: disconnect your two 1TB drives where you have your data. Second: try to fix the system when using only the 2 200GB disks. Losing the system will be bad, losing the data would be worse. When the system is OK again, you reconnect the data drive and make them available. When you have big drives/partitions, it is often a good idea to remove them from /etc/fstab before upgrading. Why should the upgrade progress have the filesystem with all my photos dangerously mounted? -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Part of hard disk disappeared after installing Fedora 12?
M. Fioretti wrote: I don't understand what's going on, because the disk is 350 GB. fdisk says the disk is 250GB and you have a 240GB sda2 partition. Disk /dev/sda: 250.1 GB, 250059350016 bytes 255 heads, 63 sectors/track, 30401 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Disk identifier: 0x000acea3 Device Boot Start End Blocks Id System /dev/sda1 * 1 26 204800 83 Linux Partition 1 does not end on cylinder boundary. /dev/sda2 26 30401 243991201 8e Linux LVM sda2 is a pv (physical volume) of around 230GB, as you can see here. [r...@polaris ~]# pvdisplay --- Physical volume --- PV Name /dev/sda2 VG Name vg_polaris PV Size 232.69 GB / not usable 673.00 KB 230/240/250 is all the same, as there are some confusing definitions of how big a gigabyte is. So, what is happening is that: - either the disk is 250GB and not 350GB (try hdparm /dev/sda, hdparm -i /dev/sda) - or the disk is 350GB but it is partitioned as if it were 250GB; this can happen if you copy an entire 250GB disk to a bigger one (dd if=dev/sda of=/dev/sdb), because you copy the partition table values which specify how big the disk is. Let's see what hdparm thinks. Model number + google search will tell us how big your disk really is. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: HP DL360 G6 ILLEGAL OPCODE
Andrew Hall wrote: Just installed Fedora 12 on an HP DL360 G6 which went on fine. But as soon as I attempt to boot the OS - at the point GRUB stage 1 should load - I get a red screen with an illegal opcode. Try to install again - all fine. Reboot again - illegal opcode. Is anyone else seeing this with HP hardware ? Something similar on a ProLiant DL320 G5p on Fedora 9. GRUB instantly reboots the machine instead of booting. I tried a lot of things, but I have never got the machine booting from the disks. I made a bootable iso on a rewriteable disc and left the CDROM always in (this is a server). Something is wrong at the BIOS/GRUB level, because the machine is perfectly stable after the boot. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: F12: ssh change port number
Soo-Hyun Choi wrote: Hooray, yes, I have overlooked the iptables settings - I thought there wasn't any firewall rules as it was just a fresh installation, but there were! I have disabled the iptables, and everything is working fine now. You should add a rule for the new port and reenable the firewall. (if you use a port != 22 you are interested in security, I suppose) -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Curiosity re the term 'Kit' ?!?
William Case wrote: Don't get me wrong, I am not objecting. In fact, I think adding 'Kit' to a bundle is descriptive and memorable. If there is a definition or at least a clear understanding of what a 'Kit' means its addition to a name can be that more informative. Fashion, started by Apple. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: cpio to ext4 seems much slower than to ext2, ext3 or xfs
Ric Wheeler wrote: In our testing with f12, I build a 60TB ext4 file system with 1 billion small files. A forced fsck of ext4 finished in 2.5 hours give or take a bit :-) The fill was artificial and the file system was not aged, so real world results will probably be slower. fsck time scales mostly with the number of allocated files in my experience. Allocated blocks (fewer very large files) are quite quick. What kind of machine did you use? With 60TB a simple allocation bitmap for 4k-blocks takes almost 2GB; and this is just to detect free space or double allocation of blocks. Wow. -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: How to Backup and Restore MBR within Logical Volumes?
Mr. Teo En Ming (Zhang Enming) wrote: # kpartx -av /dev/loop1 [...] But the problem is that I can only backup/clone the filesystems of my virtual machine within a logical volume. I can't backup the Master Boot Record (MBR) of the virtual machine within a logical volume. For example, dd if=/dev/hda of=mbr.hda bs=512 count=1 Because /dev/hda resides in a logical volume. The logical volume is a virtual harddisk for my virtual machine. Maybe I didn't really understand your setup, but... isn't your mbr simply on /dev/loop1? -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Executing ImageMagick
Marko Vojinovic wrote: On Tuesday 10 November 2009 21:17:15 Jim wrote: If it's a command line tool, then imagemagick isn't the execute command, because it won't execute. What would the command be ? Please RTFM, man ImageMagick. ImageMagick is a name for a whole suite of individual command-line tools, namely: convert, identify, composite, montage, compare, display, animate, import, conjure, quantize, miff. Also do a man on each of these to see detailed explanation for individual commands. You do know what man is and how to use it, right? If not, type man man in the command prompt (without quotes), press enter and read on. The man pages are not easy to read for a tool as complex as ImageMagick. Fortunately, there is also an impressive amount of example-based documentation: http://www.imagemagick.org/Usage/ including visual explanation of the results of the commands. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Fedora 12 has gone gold
Rahul Sundaram wrote: Hopefully Nouveau gets 3D support within a couple of releases as well. Getting rid of the biggest pain of proprietary drivers would be one giant step forward for Linux on the desktop. It would be wonderful. Sometimes you have no choice for the hardware; if you get the wrong laptop, bad luck. For my job, I got an ATI thinkpad when ATI was not open. Now that ATI is open, bad luck gave me an NVidia. I'm sure Noveau will make great progress and in the mean time I will be assigned an Intel one (the new bad ones). :-) -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: Slow Performance w/ Google Earth
Charlie McVeigh wrote: Error: API mismatch: the NVIDIA kernel module has version 190.42, but this NVIDIA driver component has version 190.25. Please make sure that the kernel module and all NVIDIA driver components have the same version. I am wondering if I have some sort of library mismatch between x86_64 and i586 versions of RPMs that are installed on my machine. Possible. I happened to have an old Nvidia libGL around and it caused a lot of crashes with Google Earth. (T61p, F10, 32bit) Any thoughts on how to troubleshoot this problem would be most appreciated. rpm -qa|grep -i nvidia locate libGL ldd `which googleearth` You have to understand what Google Earth is really running. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Executing ImageMagick
Hiisi wrote: mogrify? http://www.imagemagick.org/script/mogrify.php display? -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: su/kdesu not working
Dj YB wrote: I am sorry to have to post this problem to the list, but I got no answer in the forum, and couldn't find one elsewere. I can login as root. when trying to su to root using su or su - i get su: incorrect password following some threads I added my user to the wheel group and tried to edit /etc/sudoers using visudo with no luck on both. also tried to edit /etc/pam.d/su again with no luck. I really need a solution since many operations can't be done without su normally. Big hammer: strace -f su - It will print a lot of stuff, maybe you will find something (in the last lines near the password rejection, if you are lucky). If some PAM module is involved, you may discover which one. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: su/kdesu not working
Chris Adams wrote: Once upon a time, Roberto Ragusa m...@robertoragusa.it said: Big hammer: strace -f su - You can't strace a setuid executable. You are right. Actually the execution will still happen, but with no root authority, so it will not be useful for debugging. Seeing some output incorrectly made me think it was working. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: su/kdesu not working
Dj YB wrote: # strace -u foo -f su - wow that was incredibly long output the important line was about 1023 times repeating [pid 24172] close(3) = 0 [pid 24172] close(4) = 0 [pid 24172] close(5) = -1 (EBADF) bad file descriptor [pid 24172] close(6) = -1 (EBADF) bad file descriptor [pid 24172] close(7) = -1 (EBADF) bad file descriptor [pid 24172] close(8) = -1 (EBADF) bad file descriptor ... [pid 24172] close(1023) = -1 (EBADF) bad file descriptor that was before and after the password entering step. any idea what to do next? it seems like a filesystem problem No, that is not a significant part. This is just a new process starting and doing a let me close every file descriptor they passed me function. It actually closes only 3 and 4 as the others are not opened (closing them fails because they are simply not open). Your problem is somewhere else. Try to see what happens before the access denied message. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Errors from ata6:00 -- How to find corresponding device?
Chris Tyler wrote: What does the 6 or 6:00 correspond to? It doesn't appear to be a major or minor device number, or correspond to any entries in /sys that I can find. If the 6 is not the device (but rather the driver version or something), is there any part of these messages that indicate which device is associated with the error? Maybe you can find useful hints in /var/log/dmesg (or /var/log/messages). -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: X86 performance monitoring
Frank Cox wrote: My brother just sent me this inquiry and I thought I would ask here to see if any of you lot know anything about this before I send him a reply. Just checking whether you know anything about the various packages I could use to monitor performance of code running on X86 processors. I want to measure things like: - instructions executed, in a block of code (ie., from here to there) - cache misses (i-cache, d-cache, L1, L2, etc) - processor cycles - SSE4 instructions executed - ... and so on. While not precisely related to your question, you could have a look at valgrind, which has a simulation mode for this kind of statistics. It will not be real data, but if you just want to estimate if this functions costs more than that one and if this change is an optimization or not, it is quite good. With proper options it gives you impressing data which you can explore with kcachegrind. When examining code in this way it easily happens you discover unbelievable places where a lot of CPU is spent. The good part is that you do not have to recompile your kernel or your application. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Evince home directory
Patrick O'Callaghan wrote: On Tue, 2009-11-03 at 21:16 +0100, Roberto Ragusa wrote: I would really like that the save-as file selector would _not_ open on the obvious places, but open in the current directory (you know, it's called current for a reason). I couldn't agree more. I find this behaviour maddening. I imagine it's because when started from the GUI the concept of current directory is vague at best, but when starting from the Shell they definitely should take note of it. The GUI should acquire the cwd concept. Having the application launcher in the file explorer window instead of near the taskbar would be a good start (I'm starting this app to work here). Another idea could be to have a really quick bookmark feature on file explorer windows and then show them in file requesters for easy choice. Having it sort and/or colored by age/use would be perfect (like kate does on filenames when you have many documents open). It's also infuriating that every time I need to store a file using one of the standard file dialogues, e.g. from Firefox, I have to navigate to the right place *every time*. Even within the same session of the app, there's no short-term memory of the last place I stored a file. How can one be surprised that many people save everything on the Desktop? -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Evince home directory
Ed Greshko wrote: Besides, things are done in an inconsistent manner. For example, if I start gnome-terminal from the menus in a Gnome session it will start in $HOME. But, if I start it from the menus in a KDE session it will start in $HOME/Documents. To change the behavior I need to edit the menu item and change the Work Path since someone thought it would be a good idea to default to that directory. So we actually have: - the obvious base directory for the user ($HOME) - the obvious base directory for that kind of files ($HOME/Downloads) - the obvious preferred behavior chosen by the menu creator ($HOME/Documents) - the obvious directory chosen by the user (where he saved last time) And I actually still miss the best of all: - the current working directory If I'm in a console and type cd /mnt/documents/finance/reports/2009/11/03 oowriter I would really like that the save-as file selector would _not_ open on the obvious places, but open in the current directory (you know, it's called current for a reason). Then, when I close oowriter, I run inkscape and it should default to the cwd too. And so on. Add to this that I would really like to open nonexistent files: oowriter abc.doc when abc.doc does not exist. In this way I can just save instead of save-as. And it would be nice if the app let me save an empty document; so I can save immediately before writing anything. OK, I really had to say that. :-) -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: A question about the future of qemu package
Tom Horsley wrote: A system at work said I did have the vmx flag (actually there is a different flag for AMD chips as well), but I still couldn't use kvm. Finally discovered strange BIOS setting that would only take effect after a power cycle. So not only do you need the chip support, you need a BIOS that allows you to enable it (and the cpuinfo test won't check for that). I have heard horror stories of BIOS version which refuse to offer the ability to enable the feature (but never personally seen one), so merely checking vmx may not be enough. In fact the the kvm-intel module actually prints in the system log something like Virtualization support unavailable, disabled by BIOS. I don't know if the same applies to the AMD equivalent. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Proftpd won't execute in FC11
Jim wrote: On 10/26/2009 12:33 PM, Jim wrote: I found this in Messages; What does it mean ? Oct 26 12:36:20 localhost proftpd[15234]: localhost.localdomain - Failed binding to ::, port 21: Address already in use Oct 26 12:36:20 localhost proftpd[15234]: localhost.localdomain - Check the ServerType directive to ensure you are configured correctly. Your port 21 is already used by another program. Do you have another ftp server running? Try netstat -lntp|grep 21 to discover what it could be running. It could be vsftpd or maybe xinetd. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: PV and LVM resize ext4
Ambrogio wrote: e2fsck -f /dev/vg_taitsvolume/lv_root resize2fs -p /dev/vg_taitsvolume/lv_root 30G lvreduce --size 30G vg_taitsvolume/lv_root --test lvreduce --size 30G vg_taitsvolume/lv_root i suggest to reduce the LV to a little more than 30G, or you have to calculate at block level the right dimension. Maybe 31GB is a good choise to eliminate the risks of lost the last blocks in the filesystem. Well said. Never trust the definition of GB, the padding and rounding rules etc. Resize the filesystem to 20GB, then resize the volume to 30GB, then resize the fileystem again with no size specification: it will fill the available space (30GB). But... you can use a more simple algoritm make the LV on the new disks resize the old fs to feet the size of the new lv (a little less) dd the filesystem to the new LV resize the new fs as to feet the entire LV Agreed on this one too. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: print to PDF?
fred smith wrote: While not commandline, if you have OpenOffice.org, you can paste them into it then export as a PDF. Same thing with scribus, which is more pagination oriented than OpenOffice and has great PDF export options. (this comment coming from my experience in printing a long photo album with scribus; as I did not want to import and place the photos manually I used a script to write the scribus document, which is rather comprehensible XML) -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Gnome or KDE - why not ask?
Rahul Sundaram wrote: On 10/11/2009 07:48 PM, Patrick O'Callaghan wrote: Nevertheless, it remains true that Fedora is a Gnome-based distro! Why? Because it's all over the documentation. I doubt you'll find the phrase Fedora is a Gnome-based distro anywhere in the docs, but the fact remains that at every juncture where a DE is relevant to a system administration function, it's assumed to be Gnome unless a specific exception is made for KDE. Any specific pointers? We can fix the documentation. You are welcome to contribute. In most cases, it is a matter of not having enough people volunteering to comprehensively document every desktop environment. While not exactly documentation: http://www.fedoraproject.org/en/get-fedora Get Fedora 11 Desktop Edition Now This is the latest version of the Fedora Linux operating system featuring the GNOME desktop. And then on the right: KDE fans, go here! Have a PowerPC? Go here! Show me all download options in one page! Put in this way, it appears that KDE and PPC are just supported because there are a few weird people and weird machines around. Better replace fans with something more neutral such as users. BTW, why in the first two sentences the webmaster is talking and in the last one the web user is talking? Sounds stylistically unpolished. Suggestion: Fedora with KDE Fedora for PowerPC All Fedora download options Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Gnome or KDE - why not ask?
Marc Wilson wrote: Then step up and do the work, if you want KDE on an equal footing. It's not like the KDE people within Fedora haven't been saying this for ages and ages and ages. It became evident a few months ago that a suggestion like yours, while certainly reasonable, is not useful. Even if KDE were good/polished/complete enough (and it actually is), the proposal let the user choose desktop was replied with we do not want the user to choose. So, given: a) the user will not choose, we choose b) our choice is GNOME KDE people have been unable to change a). Attacking b) is not natural to KDE people, as KDE *is* fundamentally about _choice_ (the same doesn't apply to GNOME, which is fundamentally about _pre-cooked_). Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: How to dump the locked up program
Vincent Onelli wrote: Hello all, Is there way to dump the program stop responding, instead of do a full reboot?. What you are calling dump is probably called kill in the Unix world. And program is better spelled process. So, a simple Google search for How do I kill a process in Linux? will give you a lot of answers. In a console: kill 666 (where 666 is the PID of the process) Via GUI, it depends on GNOME, KDE, whatever you are using (it could be Ctrl-Esc or similar key commands). -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: booting fedora 11 from a fedora 10 grub?
James Allsopp wrote: root (hd1,0) kernel /vmlinuz-2.6.29.4-167.fc11.x86_64 ro root=UUID=7ba6f364-68c2-4dad-a54e-18c3464b0eb5 rhgb quiet initrd /initrd-2.6.29.4-167.fc11.x86_64.img Instead of selecting one of the grub options, go into the grub console (press 'c') and type these commands one after another. May be one of them tells you what's wrong. If not, use boot as last command and it should boot. (if the kernel starts but the initrd fails to find /, that's another kind of issue). One good thing of the grub console is that if you write root (hd and press tab, it will autocomplete with possible choices; the same for root (hd0, or kernel / or kernel /vml Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Your system is too slow
Michael Hennebry wrote: Most recently mplayer told me that my system was to slow. It made suggestions that were scrolled off the screen by many cpoies of AO: [pulse] pa_stream_get_latency() failed: Connection terminated AO: [pulse] pa_stream_write() failed: Connection terminated.4% 1312 0 As your attention is on the video driver, while these messages are related to audio, try if the problem persists with -nosound. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: smartd.conf corrupted by ... something
brian wrote: 2.6.30.5-43.fc11.i586 #1 SMP i686 athlon i386 I'm trying to diagnose/fix repeated hangups and found the following in /var/log/messages: smartd[2217]: Configuration file /etc/smartd.conf has fatal syntax errors. So, I opened the file and found this weirdness: -- snip -- 1418644 /sbin/lvm.static 3810 /sbin/ifup 841 /sbin/setsysfont 5207 /sbin/start_udev 17921 /sbin/dhclient-script 1124096 /sbin/fsck.ext3 366 /sbin/mpath_wait 912584 /sbin/multipath.static ... 5423 /usr/lib/python2.5/encodings/__init__.py 4376 /usr/lib/python2.5/encodings/__init__.pyc 14337 /usr/lib/python2.5/encodings/aliases.py 9100 /usr/lib/python2.5/encodings/aliases.pyc 1005 /usr/lib/python2.5/encodings/utf_8.py 1950 /usr/lib/python2.5/encodi -- snip -- First of all, I can see where my machine must have locked up again, in the middle of writing that last path. Secondly, what the heck is this?! This is nothing like what *should* be in here. What's been vandalising this file? There is a smartd.conf.RPMNEW file there, btw. Can anyone suggest what's happened here? The machine is unstable. One of the symptoms of system instability is having files with corrupted content. Not easy to understand the cause: could be hardware related (RAM, power supply, CPU) or software related (but I don't think you are using experimental filesystems or drivers). What about a memtest run? -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: [OT] Anybody using kernel 2.6.31
john wendel wrote: I built 2.6.31 from kernel.org and I'm using it now on this F11 box. It's working OK, but when I get a lot of disk I/O going, I see laggy, jerky mouse movement. I've never seen this on a Fedora kernel or home built 2.6.30. I don't hear any audio problems like skipping, just bad mouse behavior. Anybody have any suggestions for which kernel knobs I might tweak to fix this? Maybe related to this? http://www.computerworld.com.au/article/317416/kernel_2_6_31_speed_up_linux_desktop?rid=-219 It is described as an improvement, so maybe you are seeing something else. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Curious double click required
Timothy Murphy wrote: For some time, when reading an HTML email in KMail I have to click twice on activate ... by clicking here. I think that's been the case since Fedora 9 - before that I only had to click once. Recently, I've had to click twice on f=Leave=Shutdown to shutdown my laptop. The first time the little Shutdown window does not appear. Anyone else noticed this? Doesn't really matter, just odd. [I should say that I'm running KDE.] Wild guess: some window manager option related to click-to-activate-and-do-not-pass-click-to-the-window? This is a Macintosh-style behavior, I think. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Added hard drives; how to configure raid?
Reid Rivenburgh wrote: I guess that's overkill for my simple needs. This is just going to be a big /home drive. Thanks again for the help! You know that a RAID1 is not a backup solution, right? Your data will be better protected by running only one drive and using the second as a backup (external disk, for example). Sorry if I sound like one giving lessons around, but hearing you are going to have a 1.5TB home directory full of stuff pushes me to give you a warning now and not when the wrong rm command will have killed all your files. Friendly said. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Latest kernel update for F11?
Fernando Cassia wrote: Is there any automatically updated web page showing the latest greatest kernel version/build available for a given Fedora version, I mean, without getting into the repos?. Say, if I want to create a shell script to check the latest kernel on the repos and compare it with whatever is installed on the system, what would be the URL or place to check? Is screen scraping the index file of an http/ftp mirror disqualified because it is getting into the repos? :-) A well thought wget|grep|tail -1|cut would do the job. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: USB I/O performance
Tony Nelson wrote: On 09-08-30 20:06:22, Roberto Ragusa wrote: ... I have a really slow flash (writes at 2MiB/s), so I tried to increase the number; the trick was impossible for me, the parameter is refused when 120. Just out of curiosity, what happens with 64? 64 or 60 appears to slow it down a bit (2.5MiB/s becomes 2.2MiB/s). -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: In the news: Soon to be published, Skype back-door trojan code?
Marko Vojinovic wrote: But surely I am not the first one to have such an idea, or am I? You are not. People studied Skype: http://www.recon.cx/en/f/vskype-part1.pdf http://www.recon.cx/en/f/vskype-part2.pdf and found that it contains tons and tons of cryptography, obfuscation and countermeasures against debugging or reverse engineering. A closed source code like that and with an explicit purpose to build a crypted P2P network bypassing firewalls with every trick possible is something to be nervous about. The code constantly checks itself (this is why it uses a lot of CPU) and it decides things it should not (it was said that Intel convinced Skype to cripple multiconference on AMD CPUs to improve the reputation of Intel CPUs). I'd like a good alternative to Skype, even _without_ Skype compatibility. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: USB I/O performance
Bill Davidsen wrote: Didn't seem to do much for write speed on a USB connected hard drive, so the significant benefit seems limited to write to flash. It could be an interaction with the erasable-block size of the flash. Flash doesn't like small writes, and 120KiB could be small and inefficient for a 256 KiB flash block. I have a really slow flash (writes at 2MiB/s), so I tried to increase the number; the trick was impossible for me, the parameter is refused when 120. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Where is pulseaudio started?
stan wrote: If you are using the default Fedora setup, you have your own version of pulse started when you log in. I notice that it also can be started by programs that need its services, and that seems to be gconf-helper. Isn't autospawn activated by default in /etc/pulse/client.conf? That means you can have pulseaudio started by who knows what. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Removing one drive from a RAID 1 setup?
Philip Rhoades wrote: Why would it mess up grub? - I thought I would just need to edit /etc/fstab . . Well, if you remove /dev/sda and the other disks change names and you had the bootloader only on sda... -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: plotting large datasets
DJ Delorie wrote: I've got a need to plot value vs time data for 32 channels simultaneously. Gnuplot isn't up to the task (not enough uniqueness, even mixing lines and points, or control - the graph is just a mess). What else is there? Ideally, I'd like something I can interact with - enable/disable channels, highlight channels, change the time range, etc. The data comes from processed log files, so either file or API input is OK. Ideas? Try xmgrace. Old style GUI, unconventional controls, but incredibly powerful and exceptional output quality when printing (EPS export). Not immediately easy to use, but it is worth the time needed for learning. xmgrace : excel_graphs = latex : word -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: avidemux: trouble initialising audio device
James Bridge wrote: I am trying to use avidemux in fedora 11 but when I try to play a video I get the message trouble initializing audio device. The video plays, but silently. Prefs for sound give various options, including ALSA but not PulseAudio. I suppose that's the problem. Has anyone got it to work? If you run avidemux with the command pasuspender -- avidemux pulseaudio will be disabled until avidemux exits, so avidemux will access the ALSA driver. Even if it is not exactly what you asked, it could be useful as a workaround. You also probably gain a better video/audio sync, which is good if you are editing video stuff. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: nvidia vs suspend to disk
Chris Rouch wrote: I have tuxonice suspend working on F11 (and previously F10) with nvidia drivers. You need a patched kernel (kernel-tuxonice, available from atrpms) and the suspend process is different, but the end result is the same. Is this really reliable? Because tuxonice+nvidia didn't work reliably for me (80% success when resuming) around a year ago (on F8-F9?), so I decided to avoid hibernation and only use suspend-to-ram (and switch back to normal fedora kernels). Are you running 32 or 64 bit? How much RAM do you have? -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Kernel 2.6.29 for F10
Reindl Harald wrote: Hi Is there a reason that since longer time 2.6.29-Kernels for Fedora 10 landing in Updates-testing followed with 2.6.27-Builds in the meantime? On machines with kmods this is a real problem because yum detetects the kmod for 2.6.27-Build and wants to install the 2.6.27 kernel which conflicts with the installed one I use the 2.6.29 since months in the meantime, F11 has it since the release and i do not realize why F10 does not get him I've been personally using 2.6.29.4-75 successfully for a while (uptime currently is 32 days), so I join you in asking why there are no recent kernels for f10. On this route, why no 2.6.30? There is a driver I need which is included in = 2.6.30. These days upstream 2.6.31 is almost released... BTW -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: Recompile kernel without SMP
Paul Grinberg wrote: Roberto, Actually it does have all my changes: [j...@panther linux-2.6.29.i686]$ pwd /home/josh/rpmbuild/BUILD/kernel-2.6.29/linux-2.6.29.i686 [j...@panther linux-2.6.29.i686]$ cat config-i686-PAE | grep SMP CONFIG_BROKEN_ON_SMP=y # CONFIG_SMP is not set CONFIG_X86_FIND_SMP_CONFIG=y # CONFIG_X86_VSMP is not set CONFIG_SCSI_SAS_HOST_SMP=y CONFIG_VIDEO_VP27SMPX=m [j...@panther linux-2.6.29.i686]$ There is still something weird with what you are doing. The description of your steps is not realistic. You edit the .config inside BUILD but you have still not run anything which creates the files inside BUILD (you have no rpmbuild -bp and the src.rpm installation is missing too). The rpmbuild -bb will create the BUILD stuff, but this is after your modifications, so they can't survive. I think you are actually modifying something and building something else. If you are interested in understanding, start again from zero (rm -rf rpmbuild), do every step and paste here the entire shell session. Be investigative on the effect of every step: what has been created? did the resulting rpm have a different size from the normal SMP one? Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Recompile kernel without SMP
Paul Grinberg wrote: Roberto, I did it per instructions on the website. history | grep rpmbuild 813 rpmbuild -bb --with firmware --target=i686 kernel.spec Sorry, I reread your first message and realized that what you have done is not as wrong as it looked to me. You are copying the config back to the SOURCES dir, so the rpmbuild -bb is correct. The resulting rpm contains the config as a file named /boot/config-2.6.* Are your changes showing there? -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Recompile kernel without SMP
Paul Grinberg wrote: # Create RPM package rpmbuild -bb --with firmware --target=i686 kernel.spec Have you checked that the .config in the BUILD dir still actually contains what you want, after building? IIRC, rpmbuild -bb recreates the BUILD directory, and your changes are lost. Try editing the config files in SOURCES, instead. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Recovering from a hard X lock up
James Allsopp wrote: Hi, My Fedora 10 machine locks up X completely, but I can still login via SSH. The monitor still shows the X window, but nothing changes, no clock, mouse nothing. X processes are still running, but not sure how to go about shutting them down. Last night, I tried to get the system to shut down sanely by running shutdown -h now and although the command ran and I couldn't login via ssh anymore, the machine did not shut down. Can anyone tell me how to shutdown and restart X properly. Tried things like startx( had a x lock file exists error) and init 3, but nothing worked. Does capslock still change the LED status? If not, bad sign. You can use the ps auxww command to see the PID of the X process and then kill PID or, more strongly, kill -9 PID. (equivalenty, killall X and killall -9 X). Killing X will kill all the graphical applications and start X again if you are in runlevel 5. But, as you said that init 3 failed to shut down X, I'm afraid the hardware is left in a state where recovery is not possible. (FYI, startx -- :1 would have tried to start a second X instance, bypassing the existing lock) Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: pulseaudio 0.9.15 for F10?
Kevin Kofler wrote: Roberto Ragusa wrote: are there pulseaudio 0.9.15 rpms for F10? Scratch build here: http://koji.fedoraproject.org/koji/taskinfo?taskID=1592160 Also needs udev-extras: http://koji.fedoraproject.org/koji/buildinfo?buildID=83558 WARNING: May break more things than it fixes, YMMV. It is better than 0.9.14, but it doesn't look better than the 0.9.14+backported fix I built yesterday. There are still freezes of 2-4 seconds when mplayer starts, ends, seeks or goes in and out of paused mode. Thank you for your interest. -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
pulseaudio 0.9.15 for F10?
Hi all, are there pulseaudio 0.9.15 rpms for F10? Nothing in koji, the F11 rpms have no easily solvable dependencies, even an attempt to rebuild the src.rpm on F10 failed because it needs libtool-2.2, if I understand correctly. Is there any way to have usable sound on F10 (a currently supported version)? In particular I hope to solve the everything gets stuck problem with mplayer (and others): http://www.pulseaudio.org/ticket/440 Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: pulseaudio 0.9.15 for F10?
Lennart Poettering wrote: On Sat, 08.08.09 12:10, Roberto Ragusa (m...@robertoragusa.it) wrote: I am certainly one of those people who think that released distributions should only receive security fixes and small other bug fixes. I am one of those people who think supported distributions should receive fixes. Kudos to the KDE guys who go beyond small fixes and actually push KDE 4.3 to F10. Upgrading PA like this involves big changes and does not qualify as either security nor as small other bug fixes. If you want a newer PA I'd recommend doing the upgrade from F10 to F11. Not a good reason to upgrade. Actually I don't want a newer PA, just a working PA. So, I backported the patch myself. After applying this patch, mplayer does not get stuck, it is just the usual latency/skips/underruns... diff -urN pulseaudio-0.9.14/src/pulsecore/protocol-native.c pulseaudio-0.9.14-fix440/src/pulsecore/protocol-native.c --- pulseaudio-0.9.14/src/pulsecore/protocol-native.c 2009-01-13 00:11:38.0 +0100 +++ pulseaudio-0.9.14-fix440/src/pulsecore/protocol-native.c2009-08-08 18:46:34.0 +0200 @@ -797,7 +797,7 @@ uint32_t* prebuf, uint32_t* minreq) { -size_t frame_size; +size_t frame_size, max_prebuf; pa_usec_t orig_tlength_usec, tlength_usec, orig_minreq_usec, minreq_usec, sink_usec; pa_assert(s); @@ -923,8 +923,10 @@ if (*tlength = *minreq) *tlength = *minreq*2 + (uint32_t) frame_size; -if (*prebuf == (uint32_t) -1 || *prebuf *tlength) -*prebuf = *tlength; +max_prebuf = *tlength + (uint32_t) frame_size - *minreq; +if (*prebuf == (uint32_t) -1 || +*prebuf max_prebuf) +*prebuf = max_prebuf; } static void fix_playback_buffer_attr_post( -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: pulseaudio 0.9.15 for F10?
Roberto Ragusa wrote: After applying this patch, mplayer does not get stuck, it is just the usual latency/skips/underruns... Wrongly pasted the patch. Corrected one follows: diff -urN pulseaudio-0.9.14/src/pulsecore/memblockq.c pulseaudio-0.9.14-fix440/src/pulsecore/memblockq.c --- pulseaudio-0.9.14/src/pulsecore/memblockq.c 2009-01-13 00:10:34.0 +0100 +++ pulseaudio-0.9.14-fix440/src/pulsecore/memblockq.c 2009-08-08 18:42:28.0 +0200 @@ -90,8 +90,8 @@ pa_memblockq_set_maxlength(bq, maxlength); pa_memblockq_set_tlength(bq, tlength); -pa_memblockq_set_prebuf(bq, prebuf); pa_memblockq_set_minreq(bq, minreq); +pa_memblockq_set_prebuf(bq, prebuf); pa_memblockq_set_maxrewind(bq, maxrewind); pa_log_debug(memblockq sanitized: maxlength=%lu, tlength=%lu, base=%lu, prebuf=%lu, minreq=%lu maxrewind=%lu, @@ -782,16 +782,13 @@ if (bq-tlength bq-maxlength) pa_memblockq_set_tlength(bq, bq-maxlength); - -if (bq-prebuf bq-maxlength) -pa_memblockq_set_prebuf(bq, bq-maxlength); } void pa_memblockq_set_tlength(pa_memblockq *bq, size_t tlength) { size_t old_tlength; pa_assert(bq); -if (tlength = 0) +if (tlength = 0 || tlength == (size_t) -1) tlength = bq-maxlength; old_tlength = bq-tlength; @@ -800,49 +797,46 @@ if (bq-tlength bq-maxlength) bq-tlength = bq-maxlength; -if (bq-prebuf bq-tlength) -pa_memblockq_set_prebuf(bq, bq-tlength); - if (bq-minreq bq-tlength) pa_memblockq_set_minreq(bq, bq-tlength); +if (bq-prebuf bq-tlength+bq-base-bq-minreq) +pa_memblockq_set_prebuf(bq, bq-tlength+bq-base-bq-minreq); + bq-missing += (int64_t) bq-tlength - (int64_t) old_tlength; } +void pa_memblockq_set_minreq(pa_memblockq *bq, size_t minreq) { +pa_assert(bq); + +bq-minreq = (minreq/bq-base)*bq-base; + +if (bq-minreq bq-tlength) +bq-minreq = bq-tlength; + +if (bq-minreq bq-base) +bq-minreq = bq-base; + +if (bq-prebuf bq-tlength+bq-base-bq-minreq) +pa_memblockq_set_prebuf(bq, bq-tlength+bq-base-bq-minreq); +} + void pa_memblockq_set_prebuf(pa_memblockq *bq, size_t prebuf) { pa_assert(bq); if (prebuf == (size_t) -1) -prebuf = bq-tlength; +prebuf = bq-tlength+bq-base-bq-minreq; bq-prebuf = ((prebuf+bq-base-1)/bq-base)*bq-base; if (prebuf 0 bq-prebuf bq-base) bq-prebuf = bq-base; -if (bq-prebuf bq-tlength) -bq-prebuf = bq-tlength; +if (bq-prebuf bq-tlength+bq-base-bq-minreq) +bq-prebuf = bq-tlength+bq-base-bq-minreq; if (bq-prebuf = 0 || pa_memblockq_get_length(bq) = bq-prebuf) bq-in_prebuf = FALSE; - -if (bq-minreq bq-prebuf) -pa_memblockq_set_minreq(bq, bq-prebuf); -} - -void pa_memblockq_set_minreq(pa_memblockq *bq, size_t minreq) { -pa_assert(bq); - -bq-minreq = (minreq/bq-base)*bq-base; - -if (bq-minreq bq-tlength) -bq-minreq = bq-tlength; - -if (bq-minreq bq-prebuf) -bq-minreq = bq-prebuf; - -if (bq-minreq bq-base) -bq-minreq = bq-base; } void pa_memblockq_set_maxrewind(pa_memblockq *bq, size_t maxrewind) { diff -urN pulseaudio-0.9.14/src/pulsecore/protocol-native.c pulseaudio-0.9.14-fix440/src/pulsecore/protocol-native.c --- pulseaudio-0.9.14/src/pulsecore/protocol-native.c 2009-01-13 00:11:38.0 +0100 +++ pulseaudio-0.9.14-fix440/src/pulsecore/protocol-native.c2009-08-08 18:46:34.0 +0200 @@ -797,7 +797,7 @@ uint32_t* prebuf, uint32_t* minreq) { -size_t frame_size; +size_t frame_size, max_prebuf; pa_usec_t orig_tlength_usec, tlength_usec, orig_minreq_usec, minreq_usec, sink_usec; pa_assert(s); @@ -923,8 +923,10 @@ if (*tlength = *minreq) *tlength = *minreq*2 + (uint32_t) frame_size; -if (*prebuf == (uint32_t) -1 || *prebuf *tlength) -*prebuf = *tlength; +max_prebuf = *tlength + (uint32_t) frame_size - *minreq; +if (*prebuf == (uint32_t) -1 || +*prebuf max_prebuf) +*prebuf = max_prebuf; } static void fix_playback_buffer_attr_post( -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: Can I have a 64-bit kernel with 32-bit userland ?
john wendel wrote: Anyone have a simple recipe for building a Fedora system with a 64-bit kernel and only 32-bit applications? I'd say you have to install the 32 bit distro and then manually install the 64 bit kernel. Then you have to be a little careful with updates as yum will be confused by this kernel with wrong architecture. (never tried that, but interested in this matter) -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: RAID 1 error question - boot problem.
Robin Laing wrote: /dev/sda1 ext3 boot /dev/sda2 ext3 / raid /dev/sdb1 Swap /dev/sdb2 Unknown / raid Running mdadm /dev/sdb2 --examine shows that the partition superblock is showing RAID 1 and that it is clean. If sda2 and sdb2 are part of a RAID-1 they should be identical. The fsck should give the same result on both, but sdb2 is not even recognized as ext3. I do not know exactly the meaning of what you have pasted, but I suppose that ext3 and unknown is from content autodetection and raid is from partition type (0xfd). So you should have an md device (md0?) composed by sda2 and sdb2, which is not happening. As this is a critical system, it is a priority and is being used as a virtual server. With only the second drive installed, we tried to run fsck.ext3 on the /dev/sda2 (normally b2) with no success. We also tried /dev/md0 as Ubuntu has created the /dev/md0 from the single drive. So something is wrong on sdb2. If sda2 is good, you should just get the RAID working on one drive (degraded mode) and then readd the sdb2 to the mirror. The user has not tried to boot with only the one drive in yet. He is making a copy of the drive on a different system. Now, the question. On booting from a mirror 1 array, if there is a problem with the raid system, how does the boot process read the mdadm.conf file when it is on the RAID array that needs to be created? Is there some data that is stored in the /boot or someplace else that has the necessary info to tell the system how to build the array? It will not read mdadm.conf. The info is in the superblocks of the partitions. Is it part of the /boot/grub/device.map or /boot/System.map* ? No. Just superblocks, and possibly the initrd image you boot from. Any suggestions to where to start? Boot from a live cd and do: fdisk -l /dev/sda fdisk -l /dev/sdb blkid /dev/sda1 blkid /dev/sda2 blkid /dev/sdb1 blkid /dev/sdb2 cat /proc/mdstat dmesg|less fsck /dev/sda2 (but do not let it modify anything) fsck /dev/sdb2 (but do not let it modify anything) fsck /dev/md0(but do not let it modify anything) these will give you more info to understand what's happening. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: F11 (Thpd T30) : reproducible display loss
Beartooth wrote: On Sat, 11 Jul 2009 23:02:25 +0200, Roberto Ragusa wrote: The rest is over my head. I don't even follow why it is that ssh helps, alas! I'm supposing that the problem is in the graphics driver (X) of the graphics hardware you have in your T30. If you run the program via ssh, the hardware actually drawing things is the remote machine (not the T30), so it works. The same map program works fine on other F11 machines, under Wine. Does that not show, as I think, that the problem is specific to the T30 in some way?? My guess is an issue with the specific hardware+driver on the T30. Maybe the driver used on the T30 has a bug. Can somebody translate it down to my level, suggest how to test it, and what to do about it? I don't see an easy solution for this. You should try to change things, hoping to move the waters enough to sidestep the problem. Is the T30 updated (last rpms of wine, X and everything else)? What driver is X using? (it prints a lot of stuff in /var/log/Xorg.0.log) Maybe switching to another driver can be useful, at least as a test. Another idea: you can ssh into yourself, you know? Try that and see what happens. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Upgrade from FC6 to FC11
Tim wrote: On Fri, 2009-07-10 at 19:25 +0200, Roberto Ragusa wrote: They work perfectly and the only thing which is pushing me is the lack of SMART support Compile from source... But it's a kernel issue, IIRC. Instead of running the distro with an unofficial kernel (or an autocompiled kernel), I'd try the upgrade. I used to run compiled kernels in the 2.2 and 2.4 era, but it is now too messy with all the hal/udev dependencies. If it were a simple tool missing, I would have compiled it already. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: F11 (Thpd T30) : reproducible display loss
Beartooth wrote: [b...@hbsk2 ~]$ ssh -X 192.168.1.109 b...@192.168.1.109's password: Last login: Sun Jul 5 16:23:53 2009 from 192.168.1.102 [b...@thpd30 ~]$ env WINEPREFIX=/home/btth/.wine wine C:\Garmin\MapSource.exe fixme:bitblt:X11DRV_ClientSideDIBCopy potential optimization: pixel format conversion err:imagelist:IMAGELIST_InternalExpandBitmaps creating new image bitmap (x=23904 y=24)! fixme:bitblt:X11DRV_ClientSideDIBCopy potential optimization: pixel format conversion (Meanwhile, by comparison, PC #1 was showing these messages : [b...@hbsk2 ~]$ env WINEPREFIX=/home/btth/.wine wine C:\Garmin\MapSource.exe fixme:bitblt:X11DRV_ClientSideDIBCopy potential optimization: pixel format conversion fixme:bitblt:X11DRV_ClientSideDIBCopy potential optimization: pixel format conversion) [...] So I'm altogether at sea about what is wrong with the T30 -- the new hard drive, the old BIOS, its version of X, or what. I read your post rapidly, but I give you my two cents: the program (and/or wine) is drawing a big bitmap (possibly because of a wrong calculation about screen size etc.) and this triggers a problem in the X driver. No wonder it works when on ssh. There are only two things that can lock a machine; the kernel and X. The hard disk is not guilty, sure. :-) -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Upgrade from FC6 to FC11
Jussi Lehtola wrote: On Thu, 2009-07-09 at 23:30 +0200, Roberto Ragusa wrote: Because I have a couple of FC5 servers, and I could upgrade through the odd releases 5-7-9-11. It should be rather straightforward to upgrade them to CentOS 5.3, you should be able to do it with yum. If they're 64-bit boxes you may have to remove some .i386 packages before the transaction goes through. It is a little puzzling that when someone asks on a fedora ML how to upgrade Fedora, the answer is often upgrade to CentOS instead. Apart from the dangerous reinforcement of the Fedora is not for real stuff myth, I wonder why CentOS (or RHEL), which are based on Fedora can upgrade Fedora-CentOS when Fedora can not guarantee Fedora-Fedora. I mean, if there are scripts to convert inittab to upstart, etc., they take life in Fedora and then are included in RHEL, right? Another thing I'm not convinced about is the DVD or yum upgrade choice. Once upon a time, the DVD (the CD) was the way. Not anymore; there are proposal about removing the DVD upgrade option. So... let's use yum. But it can't do everything anaconda can, so this is at your risk too. Are we going the path to an unupgradable Fedora? In my specific case, I want to remain on Fedora as it is more familiar and more complete than CentOS. (and yes, the machines are 64 bit) -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Upgrade from FC6 to FC11
Jussi Lehtola wrote: The reason is that RHEL 5 is based on Fedora 6, so the upgrade should work out smoothly. This also means I'm almost not really upgrading anything. FC5-CentOS5 is similar to FC5-F6 (ok, with some backports, I know) Also if you haven't updated the machine so far, it's best to update to a distribution that still has many years of support to come. The software you have installed manually in FC5 should still work in CentOS 5. I have no manually installed software. One of the reason to prefer Fedora (possibly a recent one) is that the software I need has been already properly packaged with high probability. If, on the other hand, you really want to stick with Fedora, I'd do a complete reinstall since that is the quickest way to do the upgrade (doing 4 upgrades is *really* slow). Besides, that way you get a clean system and all of the advantages of the new ext4 file system. In my specific case, converting the system disks brings no advantage as they are never really accessed. As regards the data disks instead (this are fileservers in a rack), I would certainly avoid ext4 for some time for safety. The data is currently on reiserfs, whose quality is often estimated a lot worse than the actual thruth. If you want to try your luck with the successive upgrades, you still might end up with a bit of a broken system, since you will probably have to recompile the software you have installed manually. Plus, you might end up with some trouble with the upgrades themselves. I'm actually considering a 5 to 11 upgrade by DVD. I hate reinstalling. Doing it in many steps would probably not be much better. Some initrd issues, rpmsave/rpmnew cleanup etc are not a problem for me. Graphical login is almost unused, so I don't think something awful can happen for apps I'm interested in: ntp, samba, netatalk, cups. The system is on RAIDed disks, so I will disconnect one of them as an easy rollback chance in case of unresolvable issues. The machines are actually not mine and they are used in production. This is why I'm didn't upgrade in the past (except FC4-FC5). They work perfectly and the only thing which is pushing me is the lack of SMART support, which I find unforgivable for a file server (yes, yes, it is all properly backupped :-) ). Thanks to everyone giving advice. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Upgrade from FC6 to FC11
Bruno Wolff III wrote: On Fri, Jul 10, 2009 at 19:25:50 +0200, Roberto Ragusa m...@robertoragusa.it wrote: The system is on RAIDed disks, so I will disconnect one of them as an easy rollback chance in case of unresolvable issues. That's unlikely to work for you. See: https://bugzilla.redhat.com/show_bug.cgi?id=188314 I know, I know. But there are 5 disks in RAID-1, so I can mdadm --grow the array from 5 disks to 4 and upgrade. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Upgrade from FC6 to FC11
Kevin Kofler wrote: Well, you could try upgrading it step by step, like FC6-F8-F10 and then to F12 when it comes out. Skipping more than one release at a time isn't really tested or supported, so it can cause problems. But keeping the ancient release forever isn't a solution either. Are you saying that skipping one release is supported? (that is, is it supposed to work?) Because I have a couple of FC5 servers, and I could upgrade through the odd releases 5-7-9-11. It is not an easy decision. - try to upgrade directly to F11? (even if something has to be fixed manually) - try the 5,7,9,11 upgrade? (it will take a lot of time) - leave them at FC5 (they have been handling terabytes of data for years with zero issues, so why not? security is not a concern as it is a local and trusted environment) There is only one thing pushing me to upgrade: SMART is not supported by FC5 on the SATA controllers I have and these machines are *file servers*. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: new disk layout
Bill Davidsen wrote: Roberto Ragusa wrote: The decision to not save buffers and cached is debatable. Even if it is memory which can be read again from the disk, it is MUCH faster to read from the swap image in a contiguous fashion, than to seek everywhere for minutes after a suspend. That argument makes no sense for buffers, you don't want to compress them, write them to swap, read them in, decompress them, all so you can then write them to the filesystem anyway. Buffers are not always dirty data. After a sync command you don't see them drop to zero. Even if they were dirty data, it can make sense to store them sequentially and compressed, instead of writing them in the right places (which could involve many many seeks). I'm perfectly aware that they will have to be properly written after resuming, but if I say hibernate I want the machine to hibernate as soon as possible, not to force the execution of any pending activity. I would run sync before hibernation, if I wanted to write dirty buffers (in any case, suspending scripts often do sync themselves before starting). That is one of the really good things which only tuxonice has. I hope you're remembering that wrong, if it's going to be written to disk there's no benefit to doing a lot now so you can slow the restore and then do all that i/o anyway. You can make the argument for cache, at least you might save so i/o after resume. I rechecked on the net just now, and tuxonice does indeed save buffers and cached by default. I hope to see it finally merged. Don't hold your breath, for years I kept pointing out issue with suspend and being told stop whining and send patches, and I would say the patch is suspend2. It's been renamed, but still mostly not merged, and new bugs have been added. I used suspend2/tuxonice a few years ago and found it really good. The official suspend stuff never really worked. Nowadays I'm only using suspend-to-ram (damned Nvidia proprietary crap). so it's really hard to guess how much swap will actually be used. But it is really simple to decide the size of the swap partition: use a big one. There is really no disadvantage to have a bigger one. 2 GB RAM? - 5 GB swap more than enough to have 1gb swap used and suspend everything; probably large enough even when you upgrade to 4 GB RAM. I don't think anyone really cares about 2 or 3 GB of disk on a modern machine (having 2 GB of RAM makes it modern). Haven't done much embedded work? Think 512M RAM, 2GB non-volatile storage, maybe another bit of EAPROM for firmware. That's not Fedora, but it definitely is suspend/resume. But this thread was about someone with a 64 bit machine and a new 1TB disk. 2GB wasted on a 1000GB disk is not a big issue. When you run Linux on a toaster, you decide differently. :-) -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: new disk layout
Bill Davidsen wrote: In suspend to disk, see the output of the 'free' command. The data in the 'buffers' and 'cached' need not be saved, the buffers are written to the filesystem and the cached data are discarded. So the room you really need is the swap in use plus the memory in use. And it gets more complex than that, because the memory is compressed as it's written out (more for speed than size, in practice), Everything correct. The decision to not save buffers and cached is debatable. Even if it is memory which can be read again from the disk, it is MUCH faster to read from the swap image in a contiguous fashion, than to seek everywhere for minutes after a suspend. That is one of the really good things which only tuxonice has. I hope to see it finally merged. so it's really hard to guess how much swap will actually be used. But it is really simple to decide the size of the swap partition: use a big one. There is really no disadvantage to have a bigger one. 2 GB RAM? - 5 GB swap more than enough to have 1gb swap used and suspend everything; probably large enough even when you upgrade to 4 GB RAM. I don't think anyone really cares about 2 or 3 GB of disk on a modern machine (having 2 GB of RAM makes it modern). -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Arranging data on the HDD
Javier Perez wrote: Hi I am changing FC11 from a two-HDD setup to a larger one-HDD setup. The first setup is using LVM spanning both disks. What is the best way to move everything to one HDD? If you want to move the logical volumes from the disk couple to the large disk, you could add the large disk to your current LVM and then use pvmove to force the relocation of all your stuff to the large disk. After that, you can use vgreduce to remove the two disks from the volume group (from three disks to one). Also, I want to have two partitions, one for the data and one for the programs. I know it is easy to put /home on its own partition, but what about MySQL? I am using MythTV and both MySQL and MythTV use /var/lib to store their data. I'd rather not use hard links ln- s bla bla bla. I am doing that with another pc and sometimes when I had trouble with LVM the data would be written on the wrong places. Another option would be to move the data directories for MythtV and/or MySQL, but I'd hate to break something that other programs are depending on. One thing you can do is to use the bind option when mounting. You can have something in a place you like and then have it appear somewhere else too (in a place that the rpms like). For example, my system has this in /etc/fstab: /dev/sda9 /mnt/data reiserfs defaults1 3 /mnt/data/wwwoffle /var/cache/wwwoffle autobind 0 0 I'd like to hear everyone's thoughts on this, I bet I'm not the only one doing something similar, but I guess this is kind of difficult to phrase for google. Thanks -- -- /\_/\ |O O| pepeb...@gmail.com mailto:pepeb...@gmail.com Javier Perez While the night runs toward the day... m m Pepebuho watches from his high perch. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: increasing time spent on grub during pm-hibernate
Konstantin Svist wrote: And it's not necessarily limited to minor corruption - the whole filesystem could get corrupted, you could lose EVERYTHING on it. Even if you just want to boot to Windows, you usually have a few different kernel versions in grub. If you chose one of those by accident, you would get this nasty surprise, too. I think that the new kernel will simply refuse to load the hibernated image and boot normally (as after an improper shutdown). Plus, even though windows rarely mounts linux partitions, linux could have easily had the windows partition mounted (to read/write the documents). Yes. But if you do not have windows partitions mounted and you hibernate and reboot to do something in windows, grub will immediately try to resume, you understand what happens and try to shutdown everything before the kernel actually runs. If you are lucky you power on again, you manage to have to grub menu shown by pressing one key and you boot windows and then resume Linux. If you are unlucky, the resume image is not usable anymore (the system has been resumed and then powered off) and you lose all the session you wanted to preserve. And that's an additional case of inexperienced users can do stupid things, so we have to block experienced users doing smart things. (I used to run tuxonice and this holding-hands stuff was not there, luckily). Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Converting to EXT4 on encrypted lvm
Bruno Wolff III wrote: That device is the encrypted block device that lvm is stored on. tune2fs isn't going to like that though. As we are talking about the root filesystem, a simple df or mount or cat /etc/fstab could show the device mounted on the mount point /. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: accessing a samba network drive via the command line
Kevin Kempter wrote: One more question: is there a way via smbclient to execute the equivelant of a find . | wc -l ? I'm wanting a count of how many files I have on the drive, including all dirs and sub-dirs. You can mount the filesystem and use all the Unix tools you need. Something like: mkdir /mnt/raw mount 192.168.1.30:raw /mnt/raw -t cifs (a lot of possibly useful extra options in the mount manpage, such as -o guest, -o ro) Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: How do you know when a reboot is required after yum update?
Donald Russell wrote: Obviously a reboot is required when loading a new kernel But in general, does yum leave some breadcrumb clue that some other process can check? For example, maybe there's something in the package files that yum sees and creates a /rebootRequired file which is always deleted upon booting? If not, I think that'd be a cool idea.. the file could be /rebootRequired.txt and contain a plain text list of packages that caused the condition. :-) It is quite apparent that you have a Windows like mindset. I say this without any will to offend you; I just observe: 1) the reboot idea 2) the file name with lowercase and uppercase mixed 3) the file extension .txt :-) Now the serious answer. The kernel case is obvious; installing a kernel is useless without a reboot. But in other cases the new stuff will be used together with the old one. For example, replaced libs will be used by programs freshly started, and old libs will be used for programs already running. So, the reboot is not strictly required. maybe you may want to restart some services or applications. And it always depends on the specific use of the apps. Say you install a new tail command with a security related fix. Would you reboot? Maybe not. But if you have a running tail -f on some log files where things under external control may be printed? Then maybe yes. Or you just restart the logging script, right? There are some cases (glibc vulnerability for example) where a reboot could be a good idea. But, guess what, if you upgrade glibc, the sshd process will be restarted (!) as that is considered an extremely sensitive daemon. I may want to restart network facing stuff like httpd, firefox and something else, but I do not worry too much about the clock applet currently running the unfixed glibc. (The iwl Intel driver forced me to reboot my laptop after more than 150 days and I had really bad words in my mind for Intel and their buggy code; my work session was incredibly complex and useful to me... I hate reboots). -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Help: F11 anaconda doesn't see my hard drives
Sam Varshavchik wrote: See above. I can see the bloody drives just fine, from the shell on ALT-F2. Probably something different, but with an older Fedora (5?) I discovered that anaconda refuses to run RAID-1 arrays if they are in degraded mode. In my case I had a 5 disk RAID-1 root partition and I smartly decided to remove one of the disks from the array so to have an unchanged disk in case a I wanted to revert the upgrade. Anaconda refused to detect the degraded array. 4 disk in RAID-1 is not exactly degraded, I would say... :-) I had to grow the array from 5 disks to 4. I have no idea if this still applies. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Help: F11 anaconda doesn't see my hard drives
Bruno Wolff III wrote: On Tue, Jun 30, 2009 at 15:02:04 +0200, Roberto Ragusa m...@robertoragusa.it wrote: Anaconda refused to detect the degraded array. 4 disk in RAID-1 is not exactly degraded, I would say... :-) I had to grow the array from 5 disks to 4. Degraded means that all of the array elements aren't functional, not whether there is still redundancy. Ok, I know that, but I suppose that the refusal to upgrade on a degraded array would be do not do something dangerous when the array is not really in good conditions and then a (debatable) assumption that 4 mirrors out of 5 is an emergency condition. In my opinion this limitation is completely unjustified. RAID has as objective the system will go on; if an upgrade is impossible then the system is not going on, and it's happening for no real technical reason. I have no idea if this still applies. Yes. In your case you could have dropped the number of elements in the array, did the install and then bumped it back up again. If you only had two drives This is what I did. I did grow the array from 5 to 4 (that is, reverse-grow). you wouldn't have been able to, because anaconda also has a minimum number of drives for each raid type. So even a complete 1 element raid 1 array can't be used for an install. That is another unreasonable limitation. For example, mdadm is a little petulant if I try to create a 1-disk RAID-1, but it finally does it if I force it enough. These things are really annoying, as they happen in a particularly delicate moment, when you don't have your usual environment, you may not have an internet connection and you are already tired because of the slow DVD boot and other details. In the past I have patched anaconda to allow this. There have been RFEs both for allowing degraded arrays and allowing complete 1 element raid 1 arrays and both (well actually more since there are duplicates of these RFEs) have been closed won't fix. Was there a motivation for wontfix? I'm curious to listen what kind of logic has been applied. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Files corrupt on copy
Andy Campbell wrote: Hmm always out by 1 ... [trantor] ..mp2/tmp/src $while true while do while cp file1.zip new.zip ; cmp -l file1.zip new.zip while done 107757013 271 270 109383125 206 207 85093653 373 372 77726613 206 207 85093653 373 372 85899797 373 372 38258517 373 372 109383125 206 207 107757013 271 270 109383125 206 207 142459477 126 127 107757013 271 270 40550997 171 170 85093653 373 372 107757013 271 270 110261013 371 370 62581653 371 370 109383125 206 207 110526037 71 70 109383125 206 207 40550997 171 170 77726613 206 207 109383125 206 207 107757013 271 270 The error is always in the last bit. And there is also a strong similarity on the final part of the position error, if expressed in hexadecimal. If you paste your numbers into this command: $ while read a b; do printf %08x\n $a; done you get this: 066c3dd5 06850dd5 05126d15 04a20395 05126d15 051eba15 0247c755 06850dd5 066c3dd5 06850dd5 087dc255 066c3dd5 026ac255 05126d15 066c3dd5 06927315 03baeb95 06850dd5 06967e55 06850dd5 026ac255 04a20395 06850dd5 066c3dd5 The errors are all of kind: ..X5 where x=1,5,9,d. So the errors appear only at spots with a distance of 16*4=64 bytes. Strong suspicion on your hardware: CPU (defective L2 cache line?), chipset or memory (did you try one 4 GiB and then the other 4GiB stick?) Oh, hardware problems could also be caused by a defective power supply or... a motherboard not well screwed on the chassis (this one made me mad some years ago). You could also try something related to power/speed management. The only alternative is a kernel bug, but as it is touching only one bit it is not likely. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: yum 3.2.23 more than 10x slower than 3.2.21?! (BZ #508445)
Roberto Ragusa wrote: Never programmed in python, but this could be O(N^2). for pkg in pkgs[1:]: if pkg.repo.cost lowcost: msg = _('excluding for cost: %s from %s') % (pkg, pkg.repo.id) self.verbose_logger.log(logginglevels.DEBUG_3, msg) pkg.repo.sack.delPackage(pkg) Replying to myself to say a filed the problem into bugzilla. #508445 -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: yum 3.2.23 more than 10x slower than 3.2.21?!
Seth Vidal wrote: and if you can make this happen every time with 3.2.23 if you could please run: time yum -d 3 list updates | grep 'time:' You were not talking to me, but I tried and look at this: (added -C to avoid net access) # time yum -C -d 3 list updates --disablerepo=atrpms\* | grep 'time:' Config time: 0.271 repo time: 0.001 pkgsack time: 86.684 rpmdb time: 0.003 up:Obs Init time: 0.230 up:simple updates time: 0.312 up:obs time: 0.015 up:condense time: 0.000 updates time: 1.583 real1m29.527s user1m24.252s sys 0m0.409s -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: yum 3.2.23 more than 10x slower than 3.2.21?!
Seth Vidal wrote: run: time yum -d 3 --disablerepo=\* --enablerepo=ROB10\* \ --enablerepo=updates list updates | grep 'time:' and post that output. Just pasted something similar in another mail, where the time is 1:30 as there are more repositories enabled than in this test. Anyway this is it: # time yum -d 3 --disablerepo=\* --enablerepo=ROB10\* \ --enablerepo=updates list updates | grep 'time:' Config time: 0.205 repo time: 0.001 pkgsack time: 58.515 rpmdb time: 0.004 up:Obs Init time: 0.143 up:simple updates time: 0.156 up:obs time: 0.008 up:condense time: 0.000 updates time: 1.067 real1m0.595s user0m57.970s sys 0m0.519s -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: yum 3.2.23 more than 10x slower than 3.2.21?!
Seth Vidal wrote: On Fri, 26 Jun 2009, Roberto Ragusa wrote: Seth Vidal wrote: and if you can make this happen every time with 3.2.23 if you could please run: time yum -d 3 list updates | grep 'time:' You were not talking to me, but I tried and look at this: (added -C to avoid net access) # time yum -C -d 3 list updates --disablerepo=atrpms\* | grep 'time:' Config time: 0.271 repo time: 0.001 pkgsack time: 86.684 and you're POSITIVE nothing is being downloaded here? Nothing at all? Run it again, please and capture all the output and post it to a pastebin. Consider that there is a -C, that is repeatable, and that the CPU is at 100%: it is not waiting, it is computing (user time result). Just repeated the test. No disk activity. No net activity. (in any case I'm on a connection where remote downloads sometimes reach 100Mbit/s). # time yum -d 3 --disablerepo=\* --enablerepo=ROB10\* --enablerepo=updates list updates | grep 'time:' Config time: 0.199 repo time: 0.001 pkgsack time: 57.204 rpmdb time: 0.003 up:Obs Init time: 0.132 up:simple updates time: 0.152 up:obs time: 0.009 up:condense time: 0.000 updates time: 1.005 real0m59.227s user0m57.726s sys 0m0.498s -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: yum 3.2.23 more than 10x slower than 3.2.21?!
Joe Nall wrote: On Jun 26, 2009, at 10:25 AM, Roberto Ragusa wrote: Seth Vidal wrote: On Fri, 26 Jun 2009, Roberto Ragusa wrote: Seth Vidal wrote: and if you can make this happen every time with 3.2.23 if you could please run: time yum -d 3 list updates | grep 'time:' You were not talking to me, but I tried and look at this: (added -C to avoid net access) # time yum -C -d 3 list updates --disablerepo=atrpms\* | grep 'time:' Config time: 0.271 repo time: 0.001 pkgsack time: 86.684 and you're POSITIVE nothing is being downloaded here? Nothing at all? Run it again, please and capture all the output and post it to a pastebin. I have a hint and it points to the same rpms in more than one repo hypothesis. yum -d 5 lists a lot of excluding for cost: firefox-3.0.11-1.fc10.i386 from updates excluding for cost: gnome-session-2.24.3-1.fc10.i386 from updates [...] I see something related to cost handling fixes in the yum Changelog. My guess is that elements are removed from an array one at a time and everything after that is moved back one position. Or some inefficiency of that kind. You may be able to reproduce it by configuring two updates repos. The performance change occurred between 3.2.21-2 and 3.2.22-5 based on testing on koji f10 downloads. from rpm -q -changelog yum: * Tue May 12 2009 Seth Vidal skvidal at fedoraproject.org - 3.2.22-5 from ChangeLog in yum tar.gz: 2009-04-13 James Antill ja...@and.org * yum/yumRepo.py: Fix cost sorting of repos. -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: F11 upgrade - worse than Windows
Simon Andrews wrote: I saw this too (but not for that long). Going into a shell I saw that anaconda was only taking ~2%CPU and that my load averages were low and that no significant data was coming in over the network, so I'm not sure what is happening over this time. I may be wrong, but I noticed in a previous upgrade that in that phase the machine is disk-limited as it is executing a great quantity of sync() functions, related to the rpm database. I would really like a simple way to turn the sync() stuff off when doing an upgrade. I know I know, it is dangerous and libeatmydata etc., but an upgrade is already considered a risk by me, so I always backup before upgrading; at that point, I would kick the sync() stuff out of the way very happily. Best regards. -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: rpm package with many files inside
Jan Chadima wrote: Hello All I need to create rpm package with cca 50-100 tiny files inside. The whole tree is about 2-3GB binary data. Koji dies with error: Unable to create immutable header region. There are existing bug One million files means (at 4KiB per file even if its length is one byte) about 4GiB of space on most filesystems (everyone except reiserfs, IIRC). And I don't want to imagine the stress that one-million-files rpms can cause to the rpm/yum machinery, which is quite slow even in normal usage. -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
Re: Ultimate Fedora partition scheme ?
Jussi Lehtola wrote: On Thu, 2009-06-18 at 14:38 +0200, Valent Turkovic wrote: With 120GB available and with this ideas in mind would this partition scheme work or you have better ideas: sda1 ; /boot ; 200MB ; ext3 ; fedora boot sda2 ; / ; 10GB ; ext4 ; root partition for fedora sda3 ; / ; 10GB ; ext4 ; root for ubuntu (it doens't need extra /boot partition) sda5; 100GB ; extended partiton sda6; 20GB ; PV_1 for LVM (LVM physical volume) sda7; 20GB ; PV_2 for LVM sda8; 20GB ; PV_3 for LVM sda9; 20GB ; PV_4 for LVM sda10; 20GB ; PV_5 for LVM Uhh.. IMHO there is no sense whatsoever to create multiple LVM partitions on a single hard drive. You can manage with a couple small boot partitions, and the rest of stuff on LVM. When the partitions are in LVM, you can resize them whenever necessary. Let me give a different opinion on this approach. It can be useful; one day you will want to have a root for mandriva out of any LVM and you will simply kick one of the PV out of the VG and turn the partition back to native. I know there is pvresize, but I feel better when using pvmove alone (which you will probably have to use in any case before pvresize). More difficult: I want a 20 GB partition, and I'd like to have it near the fastest part of the disk (the beginning). I just kick sda6 out of the VG. Having more than one PVs has no real disadvantage, so, why not? (personally used to slice disks in 50GB or 100GB PVs) -- Roberto Ragusamail at robertoragusa.it -- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: iptables/firewall brainstorming
Thomas Woerner wrote: Roberto Ragusa wrote: //A if(port==(20-21)) PERMIT; //B if(port==(20-21) net==trusted) PERMIT; //default DENY; A wins here. The first matching rule will be used. Therefore there is no restriction for a trusted network. So your ftp server will be available for everyone - even in a public wifi. And this is exactly what it should happen. B is trying to give permissions to some machines, but it is useless, as A is giving permission to everyone. If it were: //B if(port==(20-21) net==trusted) PERMIT; //A if(port==(20-21)) PERMIT; //default DENY; then B would give permission to some machines and A would give permission to all the others, so even if the decision process is a little different the final result is the same as before. The ftp server is available for everyone. Good, so A is doing its job. :-) -- Roberto Ragusamail at robertoragusa.it -- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list