Ask Your Question

larsks's profile - activity

2019-01-13 20:22:47 -0600 received badge  Notable Question (source)
2018-06-23 18:21:29 -0600 received badge  Notable Question (source)
2018-02-12 14:53:58 -0600 received badge  Popular Question (source)
2018-02-04 12:30:24 -0600 commented question Problems installing F27 on a Macbook (installs okay, won't boot)

@cmurf any additional thoughts?

2018-02-04 12:30:08 -0600 received badge  Popular Question (source)
2017-11-26 18:09:07 -0600 commented question Problems installing F27 on a Macbook (installs okay, won't boot)

@cmurf updated with the information you requested.

2017-11-06 21:40:33 -0600 commented question Problems installing F27 on a Macbook (installs okay, won't boot)

I am trying to install only Fedora from USB media, created by downloading the install image and writing it to the flash drive with dd. The original hard drive died, so this is a replacement drive with no previous OS. I am letting Fedora create the partitions, and I get an HFS+EFI mounted at /boot/efi, an ext4 partition mounted at /boot, and an LVM VG with everything else. When the system boots, it simply doesn't recognize the hard drive (I get the flashing folder with question mark).

2017-11-06 08:56:36 -0600 asked a question Problems installing F27 on a Macbook (installs okay, won't boot)

Update 2017-11-26 Now for Fedora 27!

I have a Macbook Pro on which I am trying to install Fedora 26. The install boots and runs without a problem, but once installed the system won't boot: the Mac simply doesn't recognize the hard drive as a bootable device.

The installer creates what looks like an appropriate EFI + HFS partition.

I am trying to install only Fedora (that is, no MacOS; this is a new hard drive replacing one that had failed) from USB media, created by downloading the install image and writing it to the flash drive with dd. I am letting Fedora create the partitions, and I get an HFS+EFI mounted at /boot/efi, an ext4 partition mounted at /boot, and an LVM VG with everything else. When the system boots, it simply doesn't recognize the hard drive (I get the flashing folder with question mark).

There are some suggestions here and there about creating an MBR-style partition table instead of GPT, but (a) these are for older versions of Fedora and (b) they seem to require passing 'nogpt' on the kernel command line, which is problematic because while I can access and interact with the grub menu, the "control" key doesn't work so I can't actually execute anything after making changes to the kernel command line.

In response to @cmurf:

efibootmgr -v says:

BootCurrent: 0000
Timeout: 5 seconds
BootOrder: 0000,0080
Boot0000* Fedora    HD(1,GPT,2c1c938c-9fd7-48a4-b0c7-f73a2f2fdc5a,0x800,0x64000)/File(\EFI\fedora\shimx64.efi)
Boot0080* Mac OS X  PciRoot(0x0)/Pci(0x1f,0x2)/Sata(0,0,0)/HD(3,GPT,1aa1e829-13b7-4c37-b4be-b798f75dad4d,0x3a2500e8,0x135f20)
Boot0081* Mac OS X  PciRoot(0x0)/Pci(0x1f,0x2)/Sata(0,0,0)/HD(3,GPT,1aa1e829-13b7-4c37-b4be-b798f75dad4d,0x3a2500e8,0x135f20)
Boot0082*   PciRoot(0x0)/Pci(0x1f,0x2)/Sata(0,0,0)/HD(3,GPT,1aa1e829-13b7-4c37-b4be-b798f75dad4d,0x3a2500e8,0x135f20)
BootFFFF*   PciRoot(0x0)/Pci(0x1f,0x2)/Sata(0,0,0)/HD(2,GPT,00003353-0ee0-0000-5767-0000ca480000,0x64028,0x3a1ec0c0)/File(\System\Library\CoreServices\boot.efi)

The contents of /var/log/anaconda are available here.

The contents of /boot/efi/EFI/fedora/grub.cfg are here.

2017-10-28 08:38:30 -0600 commented question Problems with a hybrid laptop after kernel 4.13.5-200 (F26)

I'm pretty sure that's not the problem, the input-polldev module isn't loaded on my system. I've updated the question to show the lsmod difference between the working and non-working kernels.

2017-10-27 10:39:13 -0600 received badge  Good Question (source)
2017-10-26 18:27:53 -0600 received badge  Nice Question (source)
2017-10-26 08:44:35 -0600 asked a question Problems with a hybrid laptop after kernel 4.13.5-200 (F26)

I have a ThinkPad Yoga 11e running Fedora 26, which is a convertible notebook/tablet device. Prior to a recent kernel update, it worked just fine as a plain laptop, and there was no support for the tablet fcuntionality.

After upgrading to kernel 4.13.5-200, the laptop is no longer usable.

  • There appears to be support for setting the screen orientation based on the accelerometer. Unfortunately, the orientation is 90° out of sync.

  • The device is stuck in "tablet" mode, which means that the keyboard and mouse are disabled. The touch screen works and I can get an on-screen keyboard, but I'd like to use the actual keyboard.

It looks like the screen orientation issue can be solved by a simple patch to the udev hwdb. However, I'm less certain why the device thinks it's in tablet mode. In the kernel logs, I see:

[ 8.875546] thinkpad_acpi: Tablet mode switch found (type: MHKG), currently in laptop mode

...which is exactly what I saw in previous kernels (and in any case would not seem to be the source of getting stuck in tablet mode).

While it would be super awesome to have the laptop/tablet conversion working, I would prefer "just act like a laptop" over "just act like a tablet", because having no functioning keyboard and mouse makes life difficult. Where should I be looking for clues?

There is a slightly different set of modules loaded in 4.13.x vs. the working 4.11 kernel:

# diff -U0 modules.4.11.8-300.fc26.x86_64 modules.4.13.9-200.fc26.x86_64 
--- modules.4.11.8-300.fc26.x86_64  2017-10-28 09:37:43.548039245 -0400
+++ modules.4.13.9-200.fc26.x86_64  2017-10-28 09:18:57.448781996 -0400
@@ -3,0 +4,2 @@
+bmc150_accel_core
+bmc150_accel_i2c
@@ -23,0 +26 @@
+ecdh_generic
@@ -29,2 +31,0 @@
-i2c_designware_core
-i2c_designware_platform
@@ -34,0 +36,2 @@
+industrialio
+industrialio_triggered_buffer
@@ -35,0 +39 @@
+intel_int0002_vgpio
@@ -57,0 +62 @@
+kfifo_buf
@@ -138,0 +144 @@
+wmi_bmof
2016-05-03 08:23:12 -0600 commented question Changing YUM/DNF basearch on Raspberry Pi image?

You just replace $basearch in your yum repository configuration with armhfp and see if that works. I don't know enough about the ARM platform to know if those architectures are equivalent.

2016-01-10 12:33:40 -0600 received badge  Famous Question (source)
2015-08-12 21:41:44 -0600 commented question send an email with postfix from command line on fedora 22

Also, mail doesn't open an editor. Read the mail man page, and pay particular attention to the "Sending Mail" section. It's not hanging; it's waiting for you to do something.

2015-07-16 16:59:53 -0600 asked a question sssd: Using krb5 auth_provder with local id_provider?

My local username is different from my Kerberos principal. Upon logging into my system, I would like to automatically acquire a valid Kerberos ticket (assuming my local account uses the same password as my kerberos account), and I would like this ticket renewed automatically for the valid renewable lifetime of the ticket.

Is this possible using sssd?

There is a krb5_map_user option, but at least in sssd 1.12 it does not appear to be possible to combine the Kerberos auth_provider with the local id_provider.

2015-07-03 10:58:36 -0600 received badge  Notable Question (source)
2015-07-03 10:58:36 -0600 received badge  Popular Question (source)
2015-03-27 12:10:01 -0600 marked best answer Alternate site for updated live ISO images?

The /topic on #fedora points to http://alt.fedoraproject.org/pub/alt/live-respins/ for updated Live ISO images. I am getting terrible transfer rates from that site, and I was wondering if there is another source for these updates. Are they typically carried in the mirror network? I checked the torrents list at http://torrent.fedoraproject.org/ , but the ISO image there appears to be from December 2013 (the updated version at alt.fedoraproject.org dates from Jun 10).

2015-02-22 20:26:55 -0600 received badge  Critic (source)
2015-01-18 08:25:11 -0600 received badge  Great Question (source)
2015-01-18 08:25:07 -0600 received badge  Good Answer (source)
2015-01-03 20:35:51 -0600 commented answer How to start jackd with realtime scheduling under Fedora 21?

(As we discussed on IRC) That's a good suggestion, but it turns out that the ControlGroup directive was removed in systemd 205, and Fedora 21 ships with systemd 216. The best solution I've found so far is to set JoinControllers= in /etc/systemd/system.conf, which is an awfully big hammer just to get one application to run.

2015-01-02 21:34:18 -0600 answered a question fedora 21 will not extract a .zip file from command line or archive manager

The errors you are seeing suggest that your zipfile is corrupt.

If I download http://pidora.ca/pidora/releases/20/images/Pidora-2014-R3.zip , I get a file that is just over 550MB in size:

$ ls -l Pidora-2014-R3.zip 
-rw-rw-r--. 1 lars lars 576911968 Aug  1 16:32 Pidora-2014-R3.zip

WIth the following checksum:

$ openssl sha256 Pidora-2014-R3.zip 
SHA256(Pidora-2014-R3.zip)= 8a0757ae313505ebd347250566e304a1f125a4322433e5c672f04ddafc1d9cea

And unzip is able to extract that archive without a problem:

$ unzip Pidora-2014-R3.zip 
Archive:  Pidora-2014-R3.zip
  inflating: Pidora-2014-R3.img      
 extracting: Pidora-2014-R3.img.md5sum
2015-01-02 21:32:50 -0600 answered a question virt-viewer - how to find out the remote URL needed?

Assuming that you are working with QEMU libvirt domains, the documentation for the QEMU driver probably has what you are look for. For a remote TCP+SASL connection, a URL would look something like:

qemu+tcp://example.com/system

Or you can connect over ssh:

qemu+ssh://user@example.com/system
2015-01-02 21:11:47 -0600 received badge  Scholar (source)
2015-01-02 21:11:40 -0600 received badge  Popular Question (source)
2015-01-02 21:11:40 -0600 received badge  Notable Question (source)
2014-06-16 14:22:24 -0600 received badge  Self-Learner (source)
2014-06-16 14:19:21 -0600 answered a question Alternate site for updated live ISO images?

randomuser on #fedora points out that a more appropriate link for these updates is http://download.fedoraproject.org/pub/alt/live-respins , which will result in a redirect to a mirror site hosting the updated images.

In my particular case, this has transformed a > 3 hour download into a < 15 minutes download.

2014-05-28 08:22:16 -0600 received badge  Citizen Patrol (source)
2014-05-22 21:25:35 -0600 received badge  Good Question (source)
2014-05-22 21:25:35 -0600 received badge  Nice Question (source)
2014-05-21 14:15:21 -0600 received badge  Nice Answer (source)
2014-03-20 18:18:07 -0600 received badge  Famous Question (source)
2014-03-20 18:18:07 -0600 received badge  Notable Question (source)
2014-03-20 18:18:07 -0600 received badge  Popular Question (source)
2013-11-05 14:15:25 -0600 commented question I don't want to use disk UUID on Fedora 19. but how?

Why do you want to disable the use of UUIDs for mounting? This prevents problems caused by device renaming (when adding a new disk causes what used to be something like "sda" to become "sdb") or problems caused by failing to label a filesystem (when you're trying to mount LABEL=root but oops, you for to set the label). You can just go ahead and replace the UUID=... in /etc/fstab with a device path, LABEL=..., or anything else supported by mount.

2013-09-04 10:40:43 -0600 received badge  Famous Question (source)
2013-08-23 07:30:47 -0600 received badge  Notable Question (source)
2013-08-14 12:58:56 -0600 received badge  Popular Question (source)
2013-08-13 15:38:58 -0600 received badge  Popular Question (source)
2013-08-13 15:38:58 -0600 received badge  Notable Question (source)
2013-08-13 15:38:58 -0600 received badge  Famous Question (source)
2013-08-13 13:28:41 -0600 asked a question Openvpn and selinux issues

I'm trying to run OpenVPN under Fedora 19 with selinux (selinux-policy-targeted) in enforcing mode, and I'm running into an AVC I'm not entirely sure how to handle.

Starting OpenVPN from the command line as root works fine, but starting it via systemd (systemctl start openvpn@vpcbridge, where /etc/openvpn/vpcbridge.conf exists) results in:

  ERROR: Cannot ioctl TUNSETIFF tap0: Permission denied (errno=13)

And in /var/log/audit:

  type=AVC msg=audit(1376412420.435:60): avc:  denied  { relabelfrom } for
  pid=720 comm="openvpn" scontext=system_u:system_r:openvpn_t:s0
  tcontext=system_u:system_r:ifconfig_t:s0 tclass=tun_socket

For reference, here's the OpenVPN configuration:

port 1194
user openvpn
dev tap0
proto udp
secret vpcbridge.key
keepalive 10 120
persist-tun
persist-key

If I run audit2allow, I get a module file that looks like this:

  module openvpn 1.0;

  require {
        type openvpn_t;
        type ifconfig_t;
        class tun_socket relabelfrom;
  }

  #============= openvpn_t ==============
  allow openvpn_t ifconfig_t:tun_socket
  relabelfrom;

But loading that generates an error:

  # semodule -i openvpn.pp
  libsepol.print_missing_requirements: openvpn's global requirements were
  not met: type/attribute openvpn_t (No such file or directory).
  libsemanage.semanage_link_sandbox: Link packages failed (No such file or
  directory).
  semodule:  Failed!

I'm not sure what do do with this error.

2012-10-19 08:36:01 -0600 commented answer Setting default mount options for FAT filesystems

That's interesting, although solving this via udev would probably mean subverting Gnome's normal volume management mechanics. I'll have to give it a shot and see what happens.