Re: portable (really) Fedora on stick
On 23/04/2018 20:10, Samuel Sieb wrote: This makes all the initramfs files that dracut creates to be equivalent to the "rescue" version, meaning that they have all the kernel modules instead of just the ones for the system it was installed on. Does making the kernel equivalent to the "rescue" kernel also mean that all the different wireless firmwares are immediately available, for current, and historic, Eth and WiFi electronics that might be needed for when using a "portable" Fedora? I'm especially thinking about using on older laptops or PCs where various plugin cards (PCMCIA, Cardbus, EISA, PCI, which often seem to have employed widely-used electronics) might be encountered? Sounds useful, if possible. Ron ___ users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-le...@lists.fedoraproject.org
Re: using wvdial or NetworkManager while connecting on a USB 3g phone modem in India (on the Reliance GSM 3g network)
On 07/03/2016 23:07, Ranjan Maitra wrote: While using NetworkManager, the modem is detected and there pops the usual "MobileBroadband connection made which even identifies correctly the provider", but nothing actually happens (and this connection can not be edited). So I create a new Mobile Broadband connection and put in the usual entries: APN: smartnet, plan Reliance but it does not even go to connect. NetworkManager refuses connetion almost instantly and that is that. I have tried changing the phone number from the default *99# to *99***1# and even #700 though I am not sure if the latter is for CDMA connections. Either way, it made no difference so I decided to try the wvdial route. I think there are 2 different ways to use a 3G phone with Linux over USB. One way is as a sort of 'Ethernet' over USB, where the phone is treated as an automatic, 'self-deciding', device which attaches to the 3G network, and establishes its own data service (using 2G, 3G, HSPA, 4G, whatever network it can find and register on). Then, the phone will ALSO be happy to 'share' its own network connection, over USB, to and for any attached computer. In this case, the computer does not control the phone, and the USB interface is just a 480kbps connection to an Ethernet transmission device to the internet (happens to be a mobile phone), just like an ADSL modem behaves in a wired telephony example. In that case, the phone must be set up to 'share' its interconnection across the USB interface. I do not know whether that has to be set up by you on the phone, or whether the phone just does it anyway; a look through the menus might reveal something. Note that in this case Network Manager should (and, presumably, does) make no attempt to control the phone, but should allow the Fedora user to set whether either (a) the USB interface receives an IP address from the phone (ie, a DHCP arrangement with the phone as the gateway router), or (b) the USB interface is given a static IP of your own choice but, in which case, you will need to set the 'gateway' to the IP address of the phone, and you will need to set DNS server addresses to some servers that are reachable across that USB interface to the phone. For initial testing purposes, it might be simpler to allow NM to use a DHCP connection on that interface. I have sometimes seen a driver 'cdc_ether' in the logs but I am not sure whether 'cdc_ether' should be present when this mode is operating. Summarising, in this arrangement, the phone is a gateway router, attaching to the internet on its own (not under the control of the Fedora system), and the Fedora system is set to use that device as its internet gateway. The second way that an internet connection can be provided through a mobile phone is by using the phone (solely) as a modem. This is a sort of 'DUN' arrangement, as you know. I have used wvdial to do this - but not recently and not on an Android phone. If the phone is, truly, standards-compliant, then it should be possible because it should honour the relevant AT commands. I would expect it to do so. In this second arrangement the Fedora system should be set up much as you have it. The phone needs to be set up very slightly differently. The phone must not be set so as to allow itself to connect to the Internet. On my (earlier than your) Android phone, I can prevent this by turning 'off' its data service. Then, the phone's data service will be inactive, until the relevant AT command is received from Fedora, when it will attempt the connection. In this way, wvdial can control the connection, and the disconnection. I would expect that if the phone is controlling its own connection, it will likely get muddled with, or ignore, any AT commands from the Fedora system. [76131.955764] cdc_acm 3-1:1.0: ttyACM0: USB ACM device This looks good. So, I created a /etc/wvdial.conf: [Dialer LG] Init2 = ATZ Init3 = ATE0V1&D2&C1S0=0+IFC=2,2 Stupid Mode = 1 Modem Type = USB Modem ISDN = 0 Phone = *99***1# Init5 = AT+CGDCONT=1,"IP","smartnet"; Modem = /dev/ttyACM0 Username = reliance Dial Command = ATDT Password = gsm Baud = 460800 Perhaps OK. I'm slightly nervous about setting the GPRS context separately from another subsequent command with a dial string - just in case, for some reason, the action of 'dialling' triggers the use of another, 'different' context - perhaps one embedded elsewhere in the phone, or in Fedora (an earlier attempt, for example). But it may be OK. It's been a while since I did this, but I am cautious about any setup that is not explicit, and relies on anything 'implicit'. I am told that the user name and the password are not important for GSM connections as long as something is present because that is all GSM checks for. That's often the case, it seems, but I would set them to whatever Reliance expects to see. There is normally some guidance with this on the provider's help p
Structural drawing CAD pkg, with 3D view of external appearance?
List, good morning, Looking for a CAD program to draw out floorplans etc for a new building. I've access to several 2D packages but does anyone know of a package that could also display a kind-of 3D image of what the structure would finally look like? Eg, the vertical faces, roof line(s) and overall shape. We're running F23/XFCE. I had heard about, but never used or seen, a package 'sweethome3d', and there is reference to it in (maybe earlier) repos, but dnf doesn't seem to find it now. I imagine people have tried to do this, before, and I wondered if anyone could recommend anything? Grateful for any advice, regards, Ron -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
Re: Missing node0 under /sys/devices/system/node
On 24/03/2016 08:16, Kevin Wilson wrote: Any idea what can be the reason for this ? I've been a bit concerned about this, too. I get this on a notebook running F23/XFCE (upgraded from F22 using dnf). Missing node0 under /sys/devices/system/node On my m/c, this appears at F23 load time, and the m/c pauses for around 30 secs or so, then continues to start up. Everything seems to start ok so I have been letting it happen, but if there's a wider occurrence, it might be something systemic. Is it a symptom of anything problematic? (I'm offsite today but I could post additional log info etc tomorrow, if needed.) Regards, Ron -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
Re: Missing node0 under /sys/devices/system/node
On 24/03/2016 12:15, Kevin Wilson wrote: Sorry for my ignorance, what is m/c? On Thu, Mar 24, 2016 at 1:15 PM, Ron Leach wrote: Missing node0 under /sys/devices/system/node On my m/c, this appears at F23 load time, and the m/c pauses for around 30 secs or so, then continues to start up. m/c is shorthand for machine. Technical term. :) R -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
F23: Seems sshd service not accepting connections
Running F23/XFCE, with firewalld, and having commanded: # systemctl start sshd.service /var/log/secure shows that sshd is opening port 22, and listening on 0.0.0.0 . There is no entry suggesting a login attempt is received. In the XFCE GUI for firewalld, everything is running in zone labelled 'public'. - The network interface is bound to the zone 'public' - SSH service is enabled in zone 'public' - I've added port 22/tcp to the list of active ports in zone 'public' - 'Lockdown', and 'panic mode' are disabled These settings are the same in mode 'permanent' and in mode 'runtime' sshd config is set to listen on 0.0.0.0 (and ::). Our LAN uses ipv4, in the range 192.168.0/24 . This machine is connected to the network, but I cannot log in to sshd from any machine on the LAN. The client times-out after a few seconds. Clients do not report a 'rejection'. If any list readers have any ideas about what else might need to be set up, or might be going wrong, I'd be very grateful to hear, Ron -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://lists.fedoraproject.org/admin/lists/users@lists.fedoraproject.org Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
Silly mistake. Now I have two 'default routes'.
List, good evening, I wanted to change a Fedora 24 server's default outgoing route, using CLI over a local SSH connection. I found the ip command, read the man page for ip route, and gave the command: ron # ip route add default via 192.168.0.70 dev enp0s9 The server uses a static IP and its original gateway address was 192.168.0.60. Now I have ron # ip route show default via 192.168.0.70 dev enp0s9 default via 192.168.0.60 dev enp0s9 proto static metric 100 [...] So I have created two 'default' routes. I've misunderstood the man page, and I thought I'd better ask for some advice before I did any more damage. I want to reach the situation where the server has only one default route, for all traffic, which is to a gateway machine with IP address 192.168.0.70 . Should I remove both these entries, and then try 'add'ing a new default route? If so, how ought I remove these entries? Or is there a better way to correct this mistake? Grateful for any help, regards, Ron ___ users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-le...@lists.fedoraproject.org
F24 Grub chainload for Win7 fails (long)
List, good evening, Our F24 installation in dual boot mode with Windows 7 has stopped W7 from booting from the Windows menu choice in Grub's screen. F24, nevertheless, boots and operates fine. Here are the installation details. We have a new laptop with Windows 7 pre-installed, which we'd configured for use, and backed up. The BIOS is set for UEFI with legacy settings. Since W7 had occupied the whole SSD we've upgraded the SSD (cloning the W7 SSD onto the new SSD using dd) and then rechecked, and tested, W7 running on the lower 228GB of the new SSD; all was fine. We installed F24/Xfce/64 from its 'live' spin onto the the remaining (approx 250 GB) segment of the SSD, allowing it to auto-partition. Installation proceeded without error; gparted showed that F24 had created 2 more partitions above the W7 partitions. On reboot, Grub starts fine, the F24 selection boots perfectly into F24, but selecting Windows just causes the machine to 'hang', showing a constant cursor in the top left corner of the screen. Checks done so far: 1. Bios boot order now shows an additional item 'Fedora'. Switching this 'Fedora' entry (in the Bios boot list) to below the HDD boot item (in the Bios boot order) has no different effect. 2. Using the Grub CLI to set the root, the chainloader location, and give a 'boot' command, has exactly the same effect as choosing the Windows entry in the Grub menu - 'hang' with the constant cursor in the top left corner. Grub CLI shows that the Windows boot partition, and files, all seem to be in place (sda2, /EFI/Microsoft/boot/bootmgfw.efi). 3. Because some of our work activities require use of Windows, it was essential to make sure we could get back to Windows, so we ran the W7 Windows System repair CD. This identified a problem in the UEFI firmware and wanted to reset the 'fwbootmgr' item in the UEFI system (NVRAM, apparently). We let it do this - aware that it might have repercussions on the F24 installation that was working well but, since it is relatively easy to re-install Fedora, we proceeded to let Windows recover. But this had two effects: (i) Windows rebooted quite happily, without a murmur. But it rebooted only itself; Fedora's Grub menu never appeared. (ii) Though the Fedora Grub menu had gone, checking the boot order in the Bios we found that there was, now, a new entry: 'Windows Boot Loader'. This was above another entry, 'Fedora' which was never executing now because 'Windows Boot Loader' had higher priority order. Switching the boot order in the Bios - to place 'Fedora' at the top, and 'Windows Boot Loader' second - Fedora's Grub menu then appears on booting up; F24 will boot and run perfectly, but selecting the Windows choice in Grub's menus again 'hangs' (with a cursor in the the top left corner). Setting the HDD to be at the top of the boot order gave the same result as setting 'Fedora' at the top - F24 would load, but Windows would not. Conclusions: (i) UEFI seems to contain an 'order of booting'. (ii) Fedora (presumably) makes a new entry in UEFI to point to its Grub system on bootup - that aspect seems to succeed. (iii) The chainloader sequence that Fedora employs executes a Windows bootmgfw.efi (on this machine) which might be 'hanging'. (iv) Might it be possible that the F24 installer has selected a file other than that which Windows 7 itself *actually* uses when it boots? Well, apologies for the length - I'd done quite a few checks and thought it best to add those results to the post in case they help narrow things down a bit. If anyone who managed this far has any suggestions for some things to try or to check, I'd be very grateful. In particular, I'd like to check which files W7 is using in its initial boot - ie, that I could try to emulate with the Grub CLI. If I could emulate a successful W7 boot, I could edit the F24 Grub configuration and make the change permanent. Has anyone here tried something similar themselves? regards, Ron ___ users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-le...@lists.fedoraproject.org
Re: F24 Grub chainload for Win7 fails (long)
On 29/12/2016 15:46, Chris Murphy wrote: On Wed, Dec 28, 2016 at 1:55 PM, Ron Leach wrote: Grub starts fine, the F24 selection boots perfectly into F24, but selecting Windows just causes the machine to 'hang', showing a constant cursor in the top left corner of the screen. Could be this bug https://bugzilla.redhat.com/show_bug.cgi?id=1347291 Indeed it is that bug. That bug report, initially of a dual-boot problem with Windows 10, was added to with reports of similar failures with Win 8.1, and Win 7. The symptoms of the 'hang' differed slightly in each different Windows case, and the Win 7 symptoms reported there were exactly those I had experienced. While that bug was being analysed, a very useful work-around was offered, exploiting a (relatively little known but) quite flexible characteristic of the UEFI system. I'd noted that, among the various boot options listed in the BIOS, are two 'software' loaders in addition to the normally-expected hardware sources such as HDD, LAN, FDD, USB, CD, etc. Those two 'software' loaders were: 'Fedora', and 'Windows Boot Manager'. UEFI has a facility whereby if it attempts a boot from some source, including a boot from a software loader, and that boot attempt fails, UEFI will try the next source in the list. (That behaviour is common for all the hardware sources, but it also applies for the software boot loaders.) Further, UEFI is happy to let a software loader start, think, take its time, and UEFI seems willing to wait for either a successful boot, or for that software loader to 'exit' at any arbitrary time later - and then UEFI will proceed to the next entry in the boot list. Now, because of the bug the system will happily load Fedora (but not Windows) when the top boot entry is the 'Fedora' boot manager, but will happily load Windows (but not Fedora) if the top boot option is Windows Boot Manager, *and* the Fedora boot loader can be manually instructed to abort and exit, then the work-around suggested in that bug report enables the Fedora boot loader to be used, and still be able to lead to Windows being loaded. The way to achieve this is 3 steps: (i) Set the boot options to give 'Fedora' boot loader the 1st priority, and give the 'Windows Boot Manager' 2nd priority - the immediate 2nd priority. This step only has to be completed once. (ii) On boot, the Fedora Grub menu choice will be presented, with Fedora, and Windows, as selectable choices. If the user wants to run Fedora, then that choice can be selected - and Fedora will load. A further, quite simple, step is needed to load Windows. (iii) To run Windows, the Windows option should be selected on screen - but not executed yet. Instead, follow the advice at the bottom of the Grub screen to edit the Grub setting (for the selected menu choice - make sure it's Windows), by typing the letter 'e'. Grub will show its short sequence of commands that it uses to load Windows. Insert the word 'exit' into that sequence of commands - just before the line starting with 'chainloader' would be a reasonable choice. You will probably have to also hit 'return' so that 'exit' appears on its own line. That's all. Again, follow the advice at the bottom of the screen - this time to tell Grub to execute this now-altered sequence of commands - by typing 'ctrl-x'. Windows should load. This is a work-around for this bug and it can be used if a user encounters the bug and cannot (or cannot yet) install the final fix, which has, now, been fully released but has to be retrospectively applied to F24 (I am not sure whether current downloads of F25 ISOs already have this fix or not but, if not, then the work around works anyway for F25, and the fix has also been fully released for F25). Not only is this a work-around for this bug, but this trick - of letting UEFI start with one software loader, and then subsequently execute the next software loader - may be useful in other contexts as well. I think it's a useful feature of UEFI to know about. In the meantime if you have a version less than grub2-2.02-0.38 you could try updating and see if that fixes the problem. And that is the fix, finally released. I used the work-around a few times because I was slightly wary of updating Grub. Finally, I did run dnf, saw that the Grub update was version ... .38, the version described and tested in the bug report, so I did apply the update and, of course, Grub now perfectly manages this dual-boot F24 and Win 7 machine. Chris, I'm very much obliged. Ah, you had asked about another detail: On Wed, Dec 28, 2016 at 1:55 PM, Ron Leach wrote: > The BIOS is > set for UEFI with legacy settings. I don't know what this 2nd sentence means. If you can take a cell photo so th
Re: Question on the WIFI security issue Key Reinstallation Attack ("krack" attack)
On 16/10/2017 15:21, Michael Cronenworth wrote: F25: https://bodhi.fedoraproject.org/updates/FEDORA-2017-12e76e8364 F27: https://bodhi.fedoraproject.org/updates/FEDORA-2017-f45e844a85 Rawhide: (just run a dnf update) ___ Is there any longer-term security support for earlier versions? We've a few devices still running F24 or F23. If there is not a security release, is it even likely that the source for the revised versions of the affected packages might compile on F24, and might the objects even run on F24 or F23? (I imagine it is fairly unlikely, but I'd like to ask.) (We hadn't updated these devices to F25 because there had seemed to be some difficulties reported on the lists, but that would be option we still have. I've downloaded the paper to understand better the risks at (i) coffee shops etc, and (ii) whether the 'trick' can be used to gain access to a password-protected AP - such as here at our premises - in the first place. I'm concerned about both those scenarios.) Thanks for the expanded update list, regards, Ron ___ users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-le...@lists.fedoraproject.org
