nfs shares not mounted at boot
Dear list, maybe since 2 weeks (close to upgrade to Twenty_One), my nfs shares are no longer mounted at boot. Claims that the nfs server is not resolvable, but it clearly is. Raised the timeo to 200, but no luck. Hints? Bugworthy? regards Jens [root@andrea ~]# systemctl status media-jessa.mount ● media-jessa.mount - /media/jessa Loaded: loaded (/etc/fstab) Active: failed (Result: exit-code) since Sun 2015-02-22 08:52:55 CET; 1min 15s ago Where: /media/jessa What: mavie.zeeroos.int:/media/jessa Docs: man:fstab(5) man:systemd-fstab-generator(8) Process: 1849 ExecMount=/bin/mount -n mavie.zeeroos.int:/media/jessa /media/jessa -t nfs -o _netdev,rw,hard,intr,nfsvers=3,tcp,noatime,nodev,async,rsize=8192,wsize=8192,vers=3,nolock,timeo=200 (code=exited, status=32) Feb 22 08:52:55 andrea.zeeroos.de systemd[1]: media-jessa.mount mount process exited, code=exited status=32 Feb 22 08:52:55 andrea.zeeroos.de systemd[1]: Failed to mount /media/jessa. Feb 22 08:52:55 andrea.zeeroos.de systemd[1]: Unit media-jessa.mount entered failed state. Feb 22 08:52:55 andrea.zeeroos.de mount[1849]: mount.nfs: Failed to resolve server mavie.zeeroos.int: Name or service...known Hint: Some lines were ellipsized, use -l to show in full. [root@andrea ~]# host mavie.zeeroos.int mavie.zeeroos.int has address 192.168.17.124 mavie.zeeroos.int has IPv6 address 2001:6f8:11d5:0:215:17ff:fe36:aa4e -- 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: nfs shares not mounted at boot
On 02/22/2015 09:00 AM, Jens Neu wrote: Dear list, maybe since 2 weeks (close to upgrade to Twenty_One), my nfs shares are no longer mounted at boot. Claims that the nfs server is not resolvable, but it clearly is. Raised the timeo to 200, but no luck. Hints? Bugworthy? regards Jens forgot to add, mount -a after system is running works just fine: [root@andrea ~]# mount -a [root@andrea ~]# mount | grep julie mavie.zeeroos.int:/media/julie-raid on /media/julie-raid type nfs (rw,nodev,noatime,vers=3,rsize=8192,wsize=8192,namlen=255,hard,nolock,proto=tcp,timeo=200,retrans=2,sec=sys,mountaddr=192.168.17.124,mountvers=3,mountport=59740,mountproto=tcp,local_lock=all,addr=192.168.17.124,_netdev) -Jens -- 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: nfs shares not mounted at boot
On 02/22/15 16:06, Jens Neu wrote: > On 02/22/2015 09:00 AM, Jens Neu wrote: >> Dear list, >> >> maybe since 2 weeks (close to upgrade to Twenty_One), my nfs shares are >> no longer mounted at boot. Claims that the nfs server is not resolvable, >> but it clearly is. Raised the timeo to 200, but no luck. Hints? Bugworthy? >> >> regards Jens > > > forgot to add, mount -a after system is running works just fine: > > [root@andrea ~]# mount -a > > [root@andrea ~]# mount | grep julie > mavie.zeeroos.int:/media/julie-raid on /media/julie-raid type nfs > (rw,nodev,noatime,vers=3,rsize=8192,wsize=8192,namlen=255,hard,nolock,proto=tcp,timeo=200,retrans=2,sec=sys,mountaddr=192.168.17.124,mountvers=3,mountport=59740,mountproto=tcp,local_lock=all,addr=192.168.17.124,_netdev) > > -Jens Are you using the DNS server of the nfs client system? -- If you can't laugh at yourself, others will gladly oblige. -- 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: nfs shares not mounted at boot
On 02/22/2015 09:13 AM, Ed Greshko wrote: Are you using the DNS server of the nfs client system? All my systems use 192.168.17.2 as dns server: nfs-server, client(s), etc. Technically the dns server is a CentOS 6.6 KVM machine with bind9 on the nfs-server. -- 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: nfs shares not mounted at boot
On 02/22/15 17:40, Jens Neu wrote: > On 02/22/2015 09:13 AM, Ed Greshko wrote: > >> Are you using the DNS server of the nfs client system? >> > > All my systems use 192.168.17.2 as dns server: nfs-server, client(s), etc. > Technically the dns server is a CentOS 6.6 KVM machine with bind9 on the > nfs-server. > I only see this this issue when my nfs client is running as my DNS server and the resolv.conf points to "nameserver 127.0.0.1". My guess is a race condition with network and DNS server startup and I even filed a bugzilla on it a while back. I used a workaround via rc.local. I've not continued to track the bugzilla since my network configuration has changed since then. -- If you can't laugh at yourself, others will gladly oblige. -- 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: nfs shares not mounted at boot
On Sun, 22 Feb 2015 09:00:32 +0100 Jens Neu wrote: > maybe since 2 weeks (close to upgrade to Twenty_One), my nfs shares are > no longer mounted at boot. My consistent experience is that systemd has no clue when the network is "up" if by up you mean actually capable of talking to other things on the network. Thus all of the dependencies it waits on never wait long enough. I moved a slew of things to rc.local to have them restarted with different delays between them and also run a script there which keeps trying to mount all my NFS shares in a background loop till they actually mount. Only with enough junk in rc.local does my system boot reliably. -- 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: nfs shares not mounted at boot
Yes.I have noted systemd stops the network then tries to umount nfs on fedora 20. Not really a good plan. And I also did the rc.local mount as it was not mounting on boot because it tries to mount nfs before the network is live and fails. There do seem to be some significant issues around nfs and network start/stop timing. On Sun, Feb 22, 2015 at 8:14 AM, Tom Horsley wrote: > On Sun, 22 Feb 2015 09:00:32 +0100 > Jens Neu wrote: > >> maybe since 2 weeks (close to upgrade to Twenty_One), my nfs shares are >> no longer mounted at boot. > > My consistent experience is that systemd has no clue when > the network is "up" if by up you mean actually capable of > talking to other things on the network. Thus all of the > dependencies it waits on never wait long enough. I moved > a slew of things to rc.local to have them restarted with > different delays between them and also run a script there > which keeps trying to mount all my NFS shares in a background > loop till they actually mount. Only with enough junk > in rc.local does my system boot reliably. > -- > 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 -- 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: nfs shares not mounted at boot
On 02/22/2015 03:24 PM, Roger Heflin wrote: Yes.I have noted systemd stops the network then tries to umount nfs on fedora 20. Not really a good plan. Did you try f21? I share your experience on f20, but the effect "healed by itself" with f21. Actually, this really annoying deficiency of f20 was the #1 reason for me to switch to f21. Ralf -- 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
'top' fails: 'xterm-256color': unknown terminal type
When I try to run 'top' on the Fedora 21 system I am installing I get the error: 'xterm-256color': unknown terminal type 'TERM=xterm top' also fails: 'xterm': unknown terminal type. and fails the same way no matter what the terminal type is set to. I don't have a TERMIFO_DIRS env variable set nor a ~/.terminfo/ directory. Same thing also happens when I login as a newly created user. /usr/share/terminfo looks normal and all the files and directories therein are world-readable. After reading a lot of manpages I came across 'toe' which according to 'man toe' "lists all available terminal types by primary name with descriptions." In my case, 'toe' lists nothing. However, 'toe -as' "-a: report on all of the terminal databases which ncurses would search, rather than only the first one that it finds" says, --> /etc/terminfo > /usr/share/terminfo --*-: Eterm Eterm with xterm-style color... --*-: Eterm-256color Eterm with xterm 256-colors --*-: Eterm-88color Eterm with 88 colors --*-: ansiansi/pc-term compatible with... --*-: aterm AfterStep terminal and continues with list of everything in /usr/share/- terminfo/. I don't know what the "--*-" text means. I tried replacing /etc/terminfo with a symlink to /usr/share/terminfo but that made no difference. Anyone have a clue about what is going on here? -- 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
Replacing Fedora Postgresql with non-Fedora version?
Hello all, I'm doing a new install of Fedora 21 and migrating apps and services from my old Fedora 15 machine to it. I've run into the following problem... I have Postgresql-9.3 installed from the Fedora 21 yum repo in order to satisfy any packages that need postgresql. But I need to run Postgresql-9.4 so I disabled the yum postgresql startup via systemd and installed the EDB version of 9.4 into /opt/postgresql [*]. To make the 9.4 one the "system" version, I did two things 1) Replaced the postgresql-9.3 executables in /bin/ with symlinks to their counterparts in /opt/postgresql/bin/. 2) Added a file to /etc/ld.so.d that adds /opt/postgresql/lib to the load library cache and ran ldconfig. That seems to work well, I can run all the Postgresql tools and everything seems to work correctly, I get the expected 9.4 versions of the tools, etc. Except... 1) It breaks openldap. Apparently one of the libraries in /opt/postgresql/lib is also used by openldap and is not compatible. When trying to start the openldap server: slapd: symbol lookup error: slapcat: undefined symbol: ber_sockbuf_io_udp 2) Some programs still don't work. For example, a simple python cgi script using the psycopg2 module to talk to the database and run under Apache-2.4 that claims it can't find the postgresql socket "/tmp/.s.PGSQL.5432" even though the socket does exist: ~$ ls -l /tmp/.s.P* srwxrwxrwx 1 postgres postgres 0 Feb 15 15:10 /tmp/.s.PGSQL.5432= -rw--- 1 postgres postgres 46 Feb 15 15:10 /tmp/.s.PGSQL.5432.lock It runs fine interactively in a shell. Any suggestions on how to fix the above two problems or general advice of replacing a yum-installed service with an outside one? I don't want to run two separate (9.3 and 9.4) servers or to run the 9.4 server on a non-standard port, etc -- I really want anything that uses postgresql to talk to my 9.4 server by default. [*] http://www.enterprisedb.com/products-services-training/pgdownload I know about the pgdg repository but that installs postgresql directly under /usr/ which I don't like, and support for it ends roughly when support for F21 ends and I want a solution that will still work even after F21's official EOL. I was able to use the EDB Postgresql 9.1, 9.2, 9.3 and 9.4 versions on F15 (shipped with 9.0) without any problems. -- 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
F21 partitioning circus
Hi, booted from the F21 XFCE spin and tried to create four primary partitions: /boot/efi swap / /home However, this seems to be impossible. When choosing the last of the four partitions, the F21 installer automatically generates a /dev/sda5, within an extended partition (sda4). No matter what I try, it always ends up this way. WTF?? Is there something I missed? Thanks, Heinz. -- 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: 'top' fails: 'xterm-256color': unknown terminal type
On 02/22/2015 10:16 AM, Stuart McGraw wrote: When I try to run 'top' on the Fedora 21 system I am installing I get the error: 'xterm-256color': unknown terminal type Never mind... I posted another message to the list about problems I was having trying to run two versions of Postgresql and the mucking around with the system libraries I did trying to accomplish that. Turn out the the 'top' problem was a side-effect of that mucking around. -- 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: Replacing Fedora Postgresql with non-Fedora version?
On 02/22/2015 09:35 AM, Stuart McGraw wrote: Hello all, I'm doing a new install of Fedora 21 and migrating apps and services from my old Fedora 15 machine to it. I've run into the following problem... I have Postgresql-9.3 installed from the Fedora 21 yum repo in order to satisfy any packages that need postgresql. But I need to run Postgresql-9.4 so I disabled the yum postgresql startup via systemd and installed the EDB version of 9.4 into /opt/postgresql [*]. To make the 9.4 one the "system" version, I did two things 1) Replaced the postgresql-9.3 executables in /bin/ with symlinks to their counterparts in /opt/postgresql/bin/. 2) Added a file to /etc/ld.so.d that adds /opt/postgresql/lib to the load library cache and ran ldconfig. That seems to work well, I can run all the Postgresql tools and everything seems to work correctly, I get the expected 9.4 versions of the tools, etc. Except... 1) It breaks openldap. Apparently one of the libraries in /opt/postgresql/lib is also used by openldap and is not compatible. When trying to start the openldap server: slapd: symbol lookup error: slapcat: undefined symbol: ber_sockbuf_io_udp 2) Some programs still don't work. For example, a simple python cgi script using the psycopg2 module to talk to the database and run under Apache-2.4 that claims it can't find the postgresql socket "/tmp/.s.PGSQL.5432" even though the socket does exist: ~$ ls -l /tmp/.s.P* srwxrwxrwx 1 postgres postgres 0 Feb 15 15:10 /tmp/.s.PGSQL.5432= -rw--- 1 postgres postgres 46 Feb 15 15:10 /tmp/.s.PGSQL.5432.lock It runs fine interactively in a shell. Any suggestions on how to fix the above two problems or general advice of replacing a yum-installed service with an outside one? I don't want to run two separate (9.3 and 9.4) servers or to run the 9.4 server on a non-standard port, etc -- I really want anything that uses postgresql to talk to my 9.4 server by default. [*] http://www.enterprisedb.com/products-services-training/pgdownload I know about the pgdg repository but that installs postgresql directly under /usr/ which I don't like, and support for it ends roughly when support for F21 ends and I want a solution that will still work even after F21's official EOL. I was able to use the EDB Postgresql 9.1, 9.2, 9.3 and 9.4 versions on F15 (shipped with 9.0) without any problems. I recall I had run into a similar problem many years ago. It was solved by putting a shell wrapper around the apps the broke. In the shell wrapper you set up the not only the execution PATH, but also the LD_LIBRARY_PATH. Since the number of packages that can be broken might be greater than the packages you desire to be in /opt/ then you might want to not set up symlinks in /bin and /usr/bin. Instead set up the wrapper shells only for the binaries and libraries and even the man pages pathways in /bin for the specific package(s) you prefer; but be sure to name them differently so they do not clash with existing names in /bin and /usr/bin; just for starting them up from startup scripts which you are aware of. Those starter scripts will set up the env variables from the wrapper shell. This does not always work, because some programs fork and exec a fixed path in the binaries, a path that might not be what you want. -- 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: etc-shadow
On Sat, Feb 21, 2015 at 11:58 PM, Matthew Miller wrote: > On Sat, Feb 21, 2015 at 06:07:18PM -0700, Chris Murphy wrote: >> I read this: >> http://www.aychedee.com/2012/03/14/etc_shadow-password-hash-formats/ >> But Fedora doesn't have mkpasswd by default, whereas passwd seems to >> only update shadow rather than outputting to stdout. And if there's a >> salt used I can't tell how that would be referenced. > > It's generated by the crypt function in glibc — man 3 crypt, and scroll > down to the "Glibc notes" section. Although I didn't dig further, that > says that the characters in the resulting string are drawn from the set > [a-zA-Z0-9./]; I assume that it's the same number as would be found in > a sha512sum hash, except mapped to that instead of represented as a > long hexadecimal number. (If you do want to dig further, I suppose > sha512-crypt.c is the place to look.) > > If you want to generate such a string yourself, using the crypt > function seems like the easiest way (of course using the python crypt > module or whatever). That's it. Thanks! So there is a salt listed in /etc/shadow, and 5000 rounds of SHA512 are used by default according to sha512-crypt.c. The number of rounds can be changed in /etc/pam.d/passwd. Curiously, Anaconda calls authconfig to create the key, and the resulting shadow entry contains a 16 character salt. Whereas passwd uses an 8 character salt. -- Chris Murphy -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 11:12 AM, Heinz Diehl wrote: > Is there something I missed? There is only one difference in functionality between 3 primaries + 1 extended, vs 4 primaries: the former can have more partitions added without deleting any partitions. There's no actual advantage of primary partitions on linux anyway. Extlinux depends on primary partitions, but GRUB doesn't. -- Chris Murphy -- 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: F21 partitioning circus
On 22.02.2015, Chris Murphy wrote: > There's no actual advantage of primary partitions on linux anyway. > Extlinux depends on primary partitions, but GRUB doesn't. The thing is that I no longer have the freedom to do what I want when installing (unless I've missed something crucial). If there is an advantage in using other partition schemes is another question. -- 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
Issue with Fedora21 "settings" after a power failure in the middle of an update.
Hey, I am unsure how to even look at the system now. I have installed a Fedora 21 Gnome Desktop on a USB disk and while updating+upgrading the system from the basic state (57X packages) there was a power failure. After turning on the PC I have tried to turn it on it was fine and booted. I did not tried to "verify" what happen but it was in the middle of "yum update" so I tried to complete the transaction using yum.. whatever the command was. Now the system is after an undo but "settings" menu would not start. It would show the basic icon and text on the top bar for couple seconds and then will vanish and even if I am waiting 20 minutes it's still not starting. Now I want if possible to find what is the issue and maybe fix it. I need your help and recommendations. Thanks, Eliezer -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 08:08:14PM +0100, Heinz Diehl wrote: > > There's no actual advantage of primary partitions on linux anyway. > > Extlinux depends on primary partitions, but GRUB doesn't. > The thing is that I no longer have the freedom to do what I want when > installing (unless I've missed something crucial). If there is an > advantage in using other partition schemes is another question. They're actually quite related. The installer UI is intended* to present meaningful decisions, and make those choices easier and more straightforward by not necessarily offering all the possibilities when the result is effectively the same. You can, however, pre-partition your system and use those partitions, or use kickstart to partition very flexibly. * how well it succeeds _is_ another question. -- Matthew Miller Fedora Project Leader -- 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: F21 partitioning circus
On 02/22/2015 12:08 PM, Heinz Diehl wrote: On 22.02.2015, Chris Murphy wrote: There's no actual advantage of primary partitions on linux anyway. Extlinux depends on primary partitions, but GRUB doesn't. The thing is that I no longer have the freedom to do what I want when installing (unless I've missed something crucial). If there is an advantage in using other partition schemes is another question. I have a 2TB drive formatted as a single MBR partition. I guess that's just about the limit of and MBR partition size. But what if the sector size is made to be programmable and is increased at formatting time to values like 2K bytes or even 128K bytes, and then let the FS decide how to use those sectors? -- 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: F21 partitioning circus
On Sun, 22 Feb 2015 14:13:28 -0500 Matthew Miller wrote: > * how well it succeeds _is_ another question. I gave up on installing on physical hardware as soon as the new anaconda first showed up. I don't trust a single thing the hopelessly obnoxious interface shows me and have no idea what on earth it is going to do to my disks. (Being forced to click the "Done" button way before I'm done was the straw that broke the camel's back :-). Instead, I install in a virtual machine where anaconda is free to trash the virtual disks in any way it sees fit, then I copy the virtual images to partitions I create myself adjust the grub.cfg and fstab files and boot using the configfile option of a stand alone grub instance. -- 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: F21 partitioning circus
On 02/22/2015 12:25 PM, Tom Horsley wrote: On Sun, 22 Feb 2015 14:13:28 -0500 Matthew Miller wrote: * how well it succeeds _is_ another question. I gave up on installing on physical hardware as soon as the new anaconda first showed up. I don't trust a single thing the hopelessly obnoxious interface shows me and have no idea what on earth it is going to do to my disks. (Being forced to click the "Done" button way before I'm done was the straw that broke the camel's back :-). Instead, I install in a virtual machine where anaconda is free to trash the virtual disks in any way it sees fit, then I copy the virtual images to partitions I create myself adjust the grub.cfg and fstab files and boot using the configfile option of a stand alone grub instance. Seems like grub UI could be improved by providing clearer explanations for each choice and it's consequences. -- 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: F21 partitioning circus
On 02/22/2015 12:25 PM, Tom Horsley wrote: On Sun, 22 Feb 2015 14:13:28 -0500 Matthew Miller wrote: * how well it succeeds _is_ another question. I gave up on installing on physical hardware as soon as the new anaconda first showed up. I don't trust a single thing the hopelessly obnoxious interface shows me and have no idea what on earth it is going to do to my disks. (Being forced to click the "Done" button way before I'm done was the straw that broke the camel's back :-). Instead, I install in a virtual machine where anaconda is free to trash the virtual disks in any way it sees fit, then I copy the virtual images to partitions I create myself adjust the grub.cfg and fstab files and boot using the configfile option of a stand alone grub instance. Seems like the Anaconda UI could provide much clearer explanations of available choices, and the consequenes of those choices (i.e. their impact on the drives/partitions that are VISIBLE to Anaconda). -- 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: nfs shares not mounted at boot
On 02/22/2015 06:50 PM, Ralf Corsepius wrote: I share your experience on f20, but the effect "healed by itself" with f21. Actually, this really annoying deficiency of f20 was the #1 reason for me to switch to f21. Unfortunately - whatever this problem is - it does not "heal itself" with 21, since my system is on 21 with patches up to yesterday. -Jens -- 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: Replacing Fedora Postgresql with non-Fedora version?
On Sun, Feb 22, 2015 at 09:35:54 -0700, Stuart McGraw wrote: I have Postgresql-9.3 installed from the Fedora 21 yum repo in order to satisfy any packages that need postgresql. But I need to run Postgresql-9.4 so I disabled the yum postgresql startup via systemd and installed the EDB version of 9.4 into /opt/postgresql [*]. What I would is rebuild 9.4 using the source rpm from f22 in f21, then update using the generated rpms. -- 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: F21 partitioning circus
On 22.02.2015, Matthew Miller wrote: > The installer UI is intended* to present meaningful decisions, > and make those choices easier and more straightforward.. When I chose "custom partitioning", I actually chose to do things on my own, which however won't be the case. That's weird. There's the possibility for all the others to chose automatic partitioning, which will take care of those who doesn't want to fiddle around. A custom mode should be.. custom. > ..by not necessarily offering all the possibilities when > the result is effectively the same. If the result is effectively the same is something the installer or those who implement the different partitioning checks/options actually can't know. There are some corner cases where this would be impossible, and I thought this is what a "custom partitioning" is for. Btw: I noticed that not only the partitioning scheme gets altered by using an extended partition where I didn't want it to have, but also the partition numbers itself get replaced while configuring. -- 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: F21 partitioning circus
On 22.02.2015, Tom Horsley wrote: > Instead, I install in a virtual machine where anaconda is free > to trash the virtual disks in any way it sees fit, then I > copy the virtual images to partitions I create myself > adjust the grub.cfg and fstab files and boot using the > configfile option of a stand alone grub instance. That's pretty sick :-) (but I now see *perfectly* why you're doing that). -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 12:08 PM, Heinz Diehl wrote: > On 22.02.2015, Chris Murphy wrote: > >> There's no actual advantage of primary partitions on linux anyway. >> Extlinux depends on primary partitions, but GRUB doesn't. > > The thing is that I no longer have the freedom to do what I want when > installing (unless I've missed something crucial). The whole point of a GUI installer is to take away superfluous or dangerous options, not to empower users to do what they want, however they want. Every use case must be justified, and "do what I want" is not self-justifying. This comes up between sysadmins and users, engineers and consumers, all the time. The consumer says "I want X and you should get to X by doing it this way." Umm no, you want something approximately X or maybe something not even X at all, but the process by which consumers get in the vicinity of X is not at all legitimate user domain - that's for design and engineering teams to sort out. X in this case is "I want an OS installation that works" and the installer will do that if you let it. Linux OS GUI installers are all just variations on rearranging the deck chairs from CLI tools. They present the same sorts of things, just in a GUI, and still burdens the user with a lot of nonsense. What Anaconda did with new UI is break with that tradition, and emphasize final results, not the nutty esoteric details of how to get there. Where it still frustrates is how it doesn't convey this worldview very well to the user. That's a difficult problem to solve, the result is the user still thinks they're supposed to be able to manipulate partitions. We still call this "Manual Partitioning" after all, so it's really wrongly named for a UI that almost totally deemphasizes partitions. > If there is an > advantage in using other partition schemes is another question. Nope, it's directly related to installer design and behavior. Does it make sense for the installer to make it possible, let alone easy, for the user to unwittingly wedge themselves into corner? No. -- Chris Murphy -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 12:19 PM, jd1008 wrote: > I have a 2TB drive formatted as a single MBR partition. > I guess that's just about the limit of and MBR partition size. > But what if the sector size is made to be programmable > and is increased at formatting time to values like 2K bytes > or even 128K bytes, and then let the FS decide how to use > those sectors? Physical and logical sector size is fixed by the drive (in firmware). There's no changing it. I'm skeptical whether we'll ever see 4Kn (4096 byte logical and physical sector) drives being understood by (legacy) BIOS computers. The BIOS asks the drive for LBA 0 and expects to get 512 bytes, yet with 4Kn drives, LBA 0 is 4096 bytes, so instantly you have failure. Actually, it's possible the confusion even happens at POST before BIOS asks the drive for the first sector. As for fs block sizes, right now on Linux the block size can't be larger than page size, which on x86 is 4096 bytes. Btrfs gets around this somewhat with a 4KB blocksize, while defaulting to 16KB nodesize. I'm not sure of the upper limit, I've tested 64KB nodesizes. -- Chris Murphy -- 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: F21 partitioning circus
On 02/22/2015 01:11 PM, Chris Murphy wrote: On Sun, Feb 22, 2015 at 12:19 PM, jd1008 wrote: I have a 2TB drive formatted as a single MBR partition. I guess that's just about the limit of and MBR partition size. But what if the sector size is made to be programmable and is increased at formatting time to values like 2K bytes or even 128K bytes, and then let the FS decide how to use those sectors? Physical and logical sector size is fixed by the drive (in firmware). There's no changing it. I'm skeptical whether we'll ever see 4Kn (4096 byte logical and physical sector) drives being understood by (legacy) BIOS computers. The BIOS asks the drive for LBA 0 and expects to get 512 bytes, yet with 4Kn drives, LBA 0 is 4096 bytes, so instantly you have failure. Actually, it's possible the confusion even happens at POST before BIOS asks the drive for the first sector. As for fs block sizes, right now on Linux the block size can't be larger than page size, which on x86 is 4096 bytes. Btrfs gets around this somewhat with a 4KB blocksize, while defaulting to 16KB nodesize. I'm not sure of the upper limit, I've tested 64KB nodesizes. Interesting. What about FreeBSD's UFS (sometimes aka FFS - fast fs)? Does it not also allow for FS blocksize to be > than page size? -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 12:29 PM, jd1008 wrote: > Seems like grub UI could be improved by providing clearer > explanations for each choice and it's consequences. What GRUB UI? Haha. GRUB upstream is not targeted at the mortal user. It's basically a buffet of tools for distributions, who then patch the hell out of it, making them all sufficiently different they end up with all sorts of incompatibilities. GRUB upstream will invariably say you need to build their upstream version any time you report a bug, because they have no idea to what degree distro specific patches have altered upstream intended behavior. That's how awful the cooperation and agreement really is, on this very basic plumbing aspect of Linux OS's. This is a "How many ways can we boot a computer?" project, is what it comes down to. So when the goal is finding myriad ways of achieving the exact same goal, there are no resources for things like consolidated agreement, let alone polish. I think the work just on patching GRUB exceeds the work required for each distro to have their own simple bootloader if they were to have the discipline in their GUI installers to reduce the layouts and hence ways to boot the system. It's a big circular "let's make more work for ourselves" extravaganza. But then, funny enough, if it weren't for that, we maybe wouldn't have projects like Fedora Atomic which is really quite cool. -- Chris Murphy -- 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: F21 partitioning circus
On 02/22/2015 01:47 PM, Chris Murphy wrote: On Sun, Feb 22, 2015 at 12:29 PM, jd1008 wrote: Seems like grub UI could be improved by providing clearer explanations for each choice and it's consequences. What GRUB UI? Haha. GRUB upstream is not targeted at the mortal user. It's basically a buffet of tools for distributions, who then patch the hell out of it, making them all sufficiently different they end up with all sorts of incompatibilities. GRUB upstream will invariably say you need to build their upstream version any time you report a bug, because they have no idea to what degree distro specific patches have altered upstream intended behavior. That's how awful the cooperation and agreement really is, on this very basic plumbing aspect of Linux OS's. This is a "How many ways can we boot a computer?" project, is what it comes down to. So when the goal is finding myriad ways of achieving the exact same goal, there are no resources for things like consolidated agreement, let alone polish. I think the work just on patching GRUB exceeds the work required for each distro to have their own simple bootloader if they were to have the discipline in their GUI installers to reduce the layouts and hence ways to boot the system. It's a big circular "let's make more work for ourselves" extravaganza. But then, funny enough, if it weren't for that, we maybe wouldn't have projects like Fedora Atomic which is really quite cool. Sorry, I did not mean to open some other unintentional thread. I meant to say Anaconda, and Grub slipped out of my KB :) :) -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 12:56 PM, Heinz Diehl wrote: > On 22.02.2015, Matthew Miller wrote: > >> The installer UI is intended* to present meaningful decisions, >> and make those choices easier and more straightforward.. > > When I chose "custom partitioning", I actually chose to do things on my own, No, you really aren't. From the outset of using a GUI installer, you're asking for some amount of guidance. See Arch's install method, which has no installer at all, for "doing things on your own." GUI installers vary only on the scope of how much guidance you get, there is always some guidance. As I've said, Anaconda doesn't even directly let you create partitions. You're creating mount points and volumes. The partitions are entirely incidental, and done for you behind the scenes. Any notion you have of it being about partitioning is an illusion. And that illusion is perpetrated by the installer itself by calling it Manual Partitioning. > which however won't be the case. That's weird. There's the possibility for all > the others to chose automatic partitioning, which will take care of those who > doesn't want to fiddle around. A custom mode should be.. custom. I completely disagree. More custom, more flexibility, in a GUI installer, is a trap. It directly leads to unnecessary design work, coding work, maintenance work, and bugs. If it were up to me, which it obviously isn't, I'd strip out Manual Partitioning entirely, and roll some of that function into blivet-gui. And give the installer different use case options, each of which are variations on automatic partitioning. And I'd refine that, and fix those bugs, leaving the user out of it as much as possible. That's how you get polish in an installer. Case in point: BIOSBoot and EFI System partitions. The user must create these things in Manual Partitioning and that's hopelessly flawed from all perspectives except the nutty let's give the user some sense of power and control that they don't actually have nor should they. Past installer never enabled the user to create MBR gaps. There's absolutely no good reason this installer should present boot loader partitions to users now. It requires ridiculous amounts of useless knowledge. > >> ..by not necessarily offering all the possibilities when >> the result is effectively the same. > > If the result is effectively the same is something the installer or those who > implement the different partitioning checks/options actually can't know. There > are some corner cases where this would be impossible, and I thought this is > what a "custom partitioning" is for. Nope. Manual Partitioning is really just for tweaking a guided layout. It's less guidance. > Btw: I noticed that not only the partitioning scheme gets altered by using an > extended partition where I didn't want it to have, but also the partition > numbers itself get replaced while configuring. Like I keep saying, any notion the installer gives you the ability to create partitions is an illusion. If there's a case for partitions being in a certain order, for everyone, then that should be filed as a bug/RFE so that the installer always does the best thing by default. -- Chris Murphy -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 1:21 PM, jd1008 wrote: > Interesting. > What about FreeBSD's UFS (sometimes aka FFS - fast fs)? > Does it not also allow for FS blocksize to be > than page size? The block size needing to be at or smaller than the page size is a linux kernel limitation. So if FreeBSD allows it, that's probably why, they aren't using linux. Windows NTFS also has a configurable block size above 4KB, as does OS X's HFS+. Different kernels. -- Chris Murphy -- 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
Network printer out of ink weirdness
Hi everyone,
I have already found a solution to the problem I am about to discuss, I
am just posting this in case it can help someone else. I am also
wondering if there was a better way in case I missed it.
I have a Fedora 14 desktop, an F20 laptop, and 2 F21 laptops. All run
wireless except for the desktop. All all 64-bit. I have a wired network
connected Epson WF-2540 printer, using the Foundation driver
epson-inkjet-printer-201211w-1.0.0-1lsb3.2.x86_64.rpm. I can print and
scan from any machine without issues (well, thanks to some previous help
from this forum).
The other day the printer said it was low on ink. It would still print
(and look fine) so I kept using it. Maybe that was a bad idea? It
eventually said it was out of ink (black) and refused to print anything
else. Fine, I put it in Replace Cartridge mode and replaced it. I first
tried my F14 desktop system since I use it for most everything. The
printer icon still had the warning emblem on it and would not print
anything. In system-config-printer I fiddled with the Enable button and
eventually got it working again. Okay, fine, so I do the same with my F21
systems. They work as well.
Now comes the fun part: I do the same with my F20 laptop but it refuses
to print anything, I then realize it can't even ping it. All of the
other machines can ping (192.168.1.10) but not F20. F20 can ping
everything else, and is on the Net. I close down and then reactivate the
wireless interface, no good. I reboot, which I despise doing, no help
(this really pissed me off). I look at the router, no help there. I do
an strace on ping and get this at the bottom:
ioctl(1, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS,
{B38400 opost isig icanon echo ...}) = 0
ioctl(1, TIOCGWINSZ, {ws_row=33, ws_col=118, ws_xpixel=0, ws_ypixel=0}) = 0
sendmsg(3, {msg_name(16)={sa_family=AF_INET, sin_port=htons(0),
sin_addr=inet_addr("192.168.1.10")},
msg_iov(1)=[{"\10\0\34\377\16P\0\1\310
\351T\0\0\0\0Vg\6\0\0\0\0\0\20\21\22\23\24\25\26\27"..., 64}],
msg_controllen=0, msg_flags=0}, 0) = 64
recvmsg(3, 0x7fff53ff24e0, 0) = -1 EAGAIN (Resource temporarily
unavailable)
It will do this forever. I tried traceroute but it just returns the
asterisks as expected.
So, what happened here? Is this a bug somewhere or did I miss something
(again)? Or maybe this was corrected in F21. I was able to get around
this problem by switching to the wired interface. It could now ping the
printer, and then I could fiddle with the Enable setting
(system-config-printer) to get it going again. After a few test pages I
then switched back to wireless and that is now also working fine.
What did I miss? I did not do anything to the printer during this time.
Jim Lewis
--
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: F21 partitioning circus
On 02/22/2015 01:09 PM, Chris Murphy wrote: I completely disagree. More custom, more flexibility, in a GUI installer, is a trap. It directly leads to unnecessary design work, coding work, maintenance work, and bugs. I can remember when custom partitioning let you do whatever you wanted, even if it was wrong. That's probably not a Good Thing for the average user, but it would be nice if there were an Expert Mode that turned off the sanity checks but made you confirm that you knew what you were doing and accepted the risk that you might create a layout that can't work. (Making sure that if you have a /boot it's big enough and that you haven't specified separate partitions for directories that have to be on the root partition would be the only exceptions.) That way, those of us with decades of experience and highly unusual requirements can do what we need without forcing the average user to work things out without a safety net. -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 2:25 PM, Joe Zeff wrote: > On 02/22/2015 01:09 PM, Chris Murphy wrote: >> >> I completely disagree. More custom, more flexibility, in a GUI >> installer, is a trap. It directly leads to unnecessary design work, >> coding work, maintenance work, and bugs. > > > I can remember when custom partitioning let you do whatever you wanted, even > if it was wrong. That's probably not a Good Thing for the average user, but > it would be nice if there were an Expert Mode that turned off the sanity > checks but made you confirm that you knew what you were doing and accepted > the risk that you might create a layout that can't work. No, because with very few exceptions, users file bugs and complain bitterly when their crazy layout doesn't work or blows up the installer. It's not worth it. (Making sure that > if you have a /boot it's big enough and that you haven't specified separate > partitions for directories that have to be on the root partition would be > the only exceptions.) That way, those of us with decades of experience and > highly unusual requirements can do what we need without forcing the average > user to work things out without a safety net. Well you're not going to convince me that highly unusual requirements is a valid reason for someone else to do the monumental amount of work to get a GUI installer to do arbitrary things for what amounts to total edge cases. -- Chris Murphy -- 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: F21 partitioning circus
On 02/22/2015 01:37 PM, Chris Murphy wrote: Well you're not going to convince me that highly unusual requirements is a valid reason for someone else to do the monumental amount of work to get a GUI installer to do arbitrary things for what amounts to total edge cases. And yet, I used to be able to do such things, so the code must have existed. And, as far as filing bugs when they fsck things up, the response is, "You were using Expert Mode, and accepted the risks. NOTABUG." -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 2:44 PM, Joe Zeff wrote: > On 02/22/2015 01:37 PM, Chris Murphy wrote: >> >> Well you're not going to convince me that highly unusual requirements >> is a valid reason for someone else to do the monumental amount of work >> to get a GUI installer to do arbitrary things for what amounts to >> total edge cases. > > > And yet, I used to be able to do such things, so the code must have existed. Give an example, and I'll take a stab at supplying an answer. But chances are Anaconda considered such capability simply not worth development resources, even if they didn't consider it a treacherously bad idea. > And, as far as filing bugs when they fsck things up, the response is, "You > were using Expert Mode, and accepted the risks. NOTABUG." No the response is "go build your own GUI installer." The code you're talking about has to be maintained by someone, it doesn't just sit there and keep on working as everything around it changes. Basically that code broke or needed too much work to hook it up to the new user interface, so it was dropped. And even if that's not the case, the code you're referring to is python2 code, so now that anacond-blivet is moving to python3 someone would have to do that migration work. If you, who seems to care about such things so much, won't do that work, then why should anyone else do it? What you're talking about might be in-scope for blivet-gui. It definitely sounds out of scope for a GUI OS installer. Windows, OS X installers have maybe 2-3 total layouts between them. And their installers are completely, totally, bullet proof. They don't ever crash, or ask the user to create required partitions, they always succeed in their penultimate goal which is to install a bootable OS. And there are essentially zero user complaints about these installers. There's nothing at all to even complain about because they don't do anything except meet their primary requirement. Not even their developers or testers even complain about the installer, it does one thing successfully. -- Chris Murphy -- 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: F21 partitioning circus
On 02/22/2015 02:01 PM, Chris Murphy wrote: If you, who seems to care about such things so much, won't do that work, then why should anyone else do it? I haven't done any programming worth mentioning since the late '80s and never learned python. My impression was that back then, anaconda used whatever standard partitioning programs were available, rather than rolling their own. Let me ask you this: could I, if needed, boot from a Live Gpartd CD/USB, set up and format things the way I wanted and then use those existing partitions when installing Fedora? If so, that's all I'd need because what I want, mostly, is a way to use my existing layout because it's grown up piece by piece, and it's easier for me to use it than to back everything up, reformat and try to shoehorn what I have into whatever layout anaconda wants to stick me with. -- 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: F21 partitioning circus
On 22.02.2015, Chris Murphy wrote: > Windows, OS X installers have maybe 2-3 total layouts between them. > And their installers are completely, totally, bullet proof. They don't > ever crash, or ask the user to create required partitions, they always > succeed in their penultimate goal which is to install a bootable OS. Frankly, the vast majority of the users of those operating systems aren't even capable of installing them by themselves. > And there are essentially zero user complaints about these installers. > There's nothing at all to even complain about because they don't do > anything except meet their primary requirement. Not even their > developers or testers even complain about the installer, it does one > thing successfully. I see. Maintainability preceeds flexibility by reducing/eliminating user influence at the same time. While it took over 100 years in medicine to reduce "i know what's best for you" and moving towards "shared decision making", it goes the other way 'round here. Fortunately, there are still distributions which let the user have the desired influence. -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 11:39:10PM +0100, Heinz Diehl wrote: > I see. Maintainability preceeds flexibility by reducing/eliminating user > influence at the same time. While it took over 100 years in medicine to reduce > "i know what's best for you" and moving towards "shared decision making", it > goes the other way 'round here. Fortunately, there are still distributions > which let the user have the desired influence. In my opinion your comments here are a little over-strong, but, yes, fundamentally, Fedora isn't meant to be a hacker/tinkerer distribution at the user level. (If you want to get involved as a hacker/tinkerer at the _development_ level, that's awesome and please do, but that's a different thing.) User control of content and devices is part of the Fedora vision*, but as we look towards having the biggest impact in the world to bring that to everyone, there are bigger battles to fight: open and free software vs. proprietary, hardware enablement with open source drivers, ease of making derivative works, transparency and traceability of builds, etc. Those are the areas we need to fight. Chris already explained (very eloquently) the reasons the installer's partitioning interface works as it does, and the goal certainly _isn't_ to "reduce influence". It's to allow users to worry about things that matter by _just doing the right thing_ where it doesn't matter for 99.99%. If you're in that 0.01, or for whatever other reason you want it to be different, there are good mechanisms in Fedora for doing basically whatever. (Seriously, consider kickstart, even at small scale.) And if that's not enough, sure, it's a big world and there _are_ awesome hacker-focused distributions. * I mean, literally: https://fedoraproject.org/wiki/Vision_statement -- Matthew Miller Fedora Project Leader -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 3:19 PM, Joe Zeff wrote: > On 02/22/2015 02:01 PM, Chris Murphy wrote: >> >> If >> you, who seems to care about such things so much, won't do that work, >> then why should anyone else do it? > > > I haven't done any programming worth mentioning since the late '80s and > never learned python. My impression was that back then, anaconda used > whatever standard partitioning programs were available, rather than rolling > their own. ? They use parted, mdadm, lvm, grub-install/mkconfig, and mkfs. But that's not where the bulk of the code is. I don't know python either, but I can still make out some sense of the complexity involved by looking at anaconda, python-blivet (that's the bulk of the storage code), and even the new python-bytesize package will give you some idea of the complexity involved in all of this. Any GUI installer is not just some dumb wrapper for existing tools, more so with Anaconda that has a huge amount of logic wrapped into it. It's worth skimming the code. 443 lines just for iSCSI (which depends on a bunch of other code, this is just the iscsi portion), devicefactory is nearly 2000 lines. The installer is substituting/emulating a human being's logic. https://github.com/rhinstaller/anaconda https://github.com/rhinstaller/blivet > Let me ask you this: could I, if needed, boot from a Live Gpartd CD/USB, set > up and format things the way I wanted and then use those existing partitions > when installing Fedora? Yes. Matthew already mentioned that. The exception is that the installer insists on root fs being formatted by the installer. I understand why this is considered safer, but at the same time I think a fsck check (no repairing) passes without errors should permit that volume to be used. This turns into a problem if you have say, hardware raid and you need to use custom mkfs options to tune the file system to the raid. With software raid, mkfs becomes aware of the underlying geometry. This isn't guaranteed with many types of hardware raid, so custom options are needed, and we have no way to do that in the installer so instead you'll have to do this post-install with fstab mount options. -- Chris Murphy -- 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: F21 partitioning circus
On 02/22/2015 03:05 PM, Chris Murphy wrote: Yes. Matthew already mentioned that. The exception is that the installer insists on root fs being formatted by the installer. I see no problem with that. If I'm doing a clean install, that's what I want anyway, and with today's machines, the amount of time it takes isn't enough to worry about. (Or, of course, I can create the partitions ahead of time, leaving them unformatted, and let anaconda do the formatting.) -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 3:39 PM, Heinz Diehl wrote: > On 22.02.2015, Chris Murphy wrote: > >> Windows, OS X installers have maybe 2-3 total layouts between them. >> And their installers are completely, totally, bullet proof. They don't >> ever crash, or ask the user to create required partitions, they always >> succeed in their penultimate goal which is to install a bootable OS. > > Frankly, the vast majority of the users of those operating systems > aren't even capable of installing them by themselves. The users don't know these things because they don't have to know them, not the other way around. There's no benefit in them knowing such things it's not intrinsically valuable knowledge for the majority. It's sufficient that a scant minority know such things. Look at even Android and cyanogen. Look at the reinvention of all OS's for mobile devices and how much simpler things are when constraining choices. Chromebooks are in that same category. Simple. Just works. They picked a layout and stuck with it. And that's not to say the layout of my cyanogen phone is exactly simple, it uses GPT partition scheme, and has 28 partitions. (Of course that's not by my choice, I had no say.) > >> And there are essentially zero user complaints about these installers. >> There's nothing at all to even complain about because they don't do >> anything except meet their primary requirement. Not even their >> developers or testers even complain about the installer, it does one >> thing successfully. > > I see. Maintainability preceeds flexibility by reducing/eliminating user > influence at the same time. While it took over 100 years in medicine to reduce > "i know what's best for you" and moving towards "shared decision making", it > goes the other way 'round here. - No, it's called picking battles. And what you're suggesting is a false dichotomy. Shared decision making actually turns into "I want partitioning with a cherry on top, you over there, go plant me a cherry tree." - No you can do it your way by using either blivet-gui, or gparted, or CLI tools in advance if you want. - Anaconda's Manual Partitioning is still one of the single most capable installer's of the lot, even if it doesn't support your specific use case. >Fortunately, there are still distributions > which let the user have the desired influence. Right. Because the problem Linux on the desktop has is it's 1000 knobs aren't enough and users need more choice. The inhibiting factor has been, this whole time, all these years, is that users really want more f'n confusion in their life. So the other day I used Yast to do an installation. Of course they're now using Btrfs by default. But it also allows me to check a box to use LVM and I thought "oh that's almost certainly a bad idea, let's see what happens" And what I got was 18 separate LV's, formatted Btrfs, instead of one Btrfs volume and 18 subvolumes (and even the 18 subvolumes is completely pathological). So yeah, less choice should be the default in any GUI ecosystem, not more choice. And besides, why in the world should so many resources go into creating advanced storage stacks only for OS installation? Why shouldn't I have such a tool for creating a big bad ass RAID with a bunch of LV's for each department? Why should I only find this in an installer, which arguably doesn't even need that? It's not it's primary use case. And it seems there's some agreement on that front, which is how the core storage portition of Anaconda got split out into its own package, python-blivet. And now there's blivet-gui which uses it, with a gparted-like UI. And OpenLMI uses python-blivet. And maybe one day soon, Cockpit on Fedora Server, will use it rather than everyone having to reinvent this wheel. -- Chris Murphy -- 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: etc-shadow
Hi, On 02/22/2015 01:23 PM, Chris Murphy wrote: On Sat, Feb 21, 2015 at 11:58 PM, Matthew Miller mailto:mat...@fedoraproject.org>> wrote: > On Sat, Feb 21, 2015 at 06:07:18PM -0700, Chris Murphy wrote: >> I read this: >> http://www.aychedee.com/2012/03/14/etc_shadow-password-hash-formats/ >> But Fedora doesn't have mkpasswd by default, whereas passwd seems to >> only update shadow rather than outputting to stdout. And if there's a >> salt used I can't tell how that would be referenced. > > It's generated by the crypt function in glibc — man 3 crypt, and scroll > down to the "Glibc notes" section. Although I didn't dig further, that > says that the characters in the resulting string are drawn from the set > [a-zA-Z0-9./]; I assume that it's the same number as would be found in > a sha512sum hash, except mapped to that instead of represented as a > long hexadecimal number. (If you do want to dig further, I suppose > sha512-crypt.c is the place to look.) > > If you want to generate such a string yourself, using the crypt > function seems like the easiest way (of course using the python crypt > module or whatever). That's it. Thanks! So there is a salt listed in /etc/shadow, and 5000 rounds of SHA512 are used by default according to sha512-crypt.c. The number of rounds can be changed in /etc/pam.d/passwd. Curiously, Anaconda calls authconfig to create the key, and the resulting shadow entry contains a 16 character salt. Whereas passwd uses an 8 character salt. Do you happen to know if there's a pre-built version of John-the-Ripper or another password testing program that's available and works with these new passwords? Thanks, Alex -- 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: F21 partitioning circus
Hi, On 02/22/2015 06:43 PM, Chris Murphy wrote: On Sun, Feb 22, 2015 at 3:39 PM, Heinz Diehl wrote: On 22.02.2015, Chris Murphy wrote: Windows, OS X installers have maybe 2-3 total layouts between them. And their installers are completely, totally, bullet proof. They don't ever crash, or ask the user to create required partitions, they always succeed in their penultimate goal which is to install a bootable OS. Frankly, the vast majority of the users of those operating systems aren't even capable of installing them by themselves. The users don't know these things because they don't have to know them, not the other way around. There's no benefit in them knowing such things it's not intrinsically valuable knowledge for the majority. It's sufficient that a scant minority know such things. Look at even Android and cyanogen. Look at the reinvention of all OS's for mobile devices and how much simpler things are when constraining choices. Chromebooks are in that same category. Simple. Just works. They picked a layout and stuck with it. And that's not to say the layout of my cyanogen phone is exactly simple, it uses GPT partition scheme, and has 28 partitions. (Of course that's not by my choice, I had no say.) On a somewhat-related note, is it now possible with F21 to create a RAID1 /boot? I can see this as being one reason for an "escape to parted/fdisk" option. I'm curious why this option has been so elusive for anaconda over the years? A situation where a failed /dev/sda in an otherwise RAID5 system is really unfortunate and requires a whole lot of extra work when things go bad. Thanks, Alex -- 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: F21 partitioning circus
On Sun, Feb 22, 2015 at 4:18 PM, Joe Zeff wrote: > On 02/22/2015 03:05 PM, Chris Murphy wrote: >> >> Yes. Matthew already mentioned that. The exception is that the >> installer insists on root fs being formatted by the installer. > > > I see no problem with that. If I'm doing a clean install, that's what I > want anyway, and with today's machines, the amount of time it takes isn't > enough to worry about. (Or, of course, I can create the partitions ahead of > time, leaving them unformatted, and let anaconda do the formatting.) One exception is root on Btrfs doesn't require a reformat, but it does require a new subvolume (which are unique fs trees, with their own inode list). And for everyone, there are certainly bugs and some bad design in the installer. This is not intentional but it's inevitable. That's why I translate "more flexibility" into "aha! more bugs!" Every feature has to justify the ensuing madness it'll create. So file bugs, make your case concisely (although verbose how to reproduce steps are best), and feel free to cc me in the bug report. It really isn't likely, or as likely, to get polish without users pointing out the warts. And idea that's been floated for future work is getting blivet-gui created layouts to more or less "plug-in" to Anaconda with minimal user intervention. So you never know, you might get to have your cake and eat it too. -- Chris Murphy -- 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: etc-shadow
On Sun, Feb 22, 2015 at 5:01 PM, Alex Regan wrote: > Do you happen to know if there's a pre-built version of John-the-Ripper or > another password testing program that's available and works with these new > passwords? I don't know that this is all that new. It's also self-describing, the /etc/shadow entry contains the 8 or 16 character salt, so whatever's doing the password+salt+SHA512rounds work knows from /etc/shadow what to do. And if rounds= is used in /etc/pam.d/passwd, the next time you use passwd, it writes out $rounds=x$ in the 2nd field in /etc/shadow, so that too is self describing. I'd expect that jtr can directly parse these variants in /etc/passwd, but I haven't tested it. -- Chris Murphy -- 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: nfs shares not mounted at boot
On Sun, 2015-02-22 at 09:14 -0500, Tom Horsley wrote: > My consistent experience is that systemd has no clue when > the network is "up" if by up you mean actually capable of > talking to other things on the network. Thus all of the > dependencies it waits on never wait long enough. There was a thread about that, last year, I think. Another target was added to solve that stupid dependency. I can't remember what it was called, but it meant "actually on-line," as opposed to "somewhere there is some aspect of a network." -- tim@localhost ~]$ uname -rsvp Linux 3.18.6-100.fc20.i686 #1 SMP Fri Feb 6 23:32:01 UTC 2015 i686 All mail to my mailbox is automatically deleted, there is no point trying to privately email me, I will only read messages posted to the public lists. George Orwell's '1984' was supposed to be a warning against tyranny, not a set of instructions for supposedly democratic governments. -- 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
tab completion FC 21
I have just upgraded a box from fc20 to fc21 using fedup and tab completion has gone bezerk. I have erased the bash-completion package but the weird behaviour continues. It is even guessing/remembering passwords. In most cases, hitting tab during command entry does nothing. In others (randomly so far as I can see) it pops up a list of words (including spelling errors). Even worse is when I try to type a path eg /usr/local/bin/foo, it puts spaces between each component and I have to backspace to remove them. How can I really get rid of this rubbish and return to the good old, simple completion? Cheers and thanks, Stephen -- 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: nfs shares not mounted at boot
On Feb 22, 2015 1:01 AM, "Jens Neu" wrote: > > Dear list, > > maybe since 2 weeks (close to upgrade to Twenty_One), my nfs shares are no longer mounted at boot. Claims that the nfs server is not resolvable, but it clearly is. Raised the timeo to 200, but no luck. Hints? Bugworthy? > > regards Jens > > [root@andrea ~]# systemctl status media-jessa.mount > ● media-jessa.mount - /media/jessa >Loaded: loaded (/etc/fstab) >Active: failed (Result: exit-code) since Sun 2015-02-22 08:52:55 CET; 1min 15s ago > Where: /media/jessa > What: mavie.zeeroos.int:/media/jessa > Docs: man:fstab(5) >man:systemd-fstab-generator(8) > Process: 1849 ExecMount=/bin/mount -n mavie.zeeroos.int:/media/jessa /media/jessa -t nfs -o _netdev,rw,hard,intr,nfsvers=3,tcp,noatime,nodev,async,rsize=8192,wsize=8192,vers=3,nolock,timeo=200 (code=exited, status=32) > > Feb 22 08:52:55 andrea.zeeroos.de systemd[1]: media-jessa.mount mount process exited, code=exited status=32 > Feb 22 08:52:55 andrea.zeeroos.de systemd[1]: Failed to mount /media/jessa. > Feb 22 08:52:55 andrea.zeeroos.de systemd[1]: Unit media-jessa.mount entered failed state. > Feb 22 08:52:55 andrea.zeeroos.de mount[1849]: mount.nfs: Failed to resolve server mavie.zeeroos.int: Name or service...known > Hint: Some lines were ellipsized, use -l to show in full. > > > [root@andrea ~]# host mavie.zeeroos.int > mavie.zeeroos.int has address 192.168.17.124 > mavie.zeeroos.int has IPv6 address 2001:6f8:11d5:0:215:17ff:fe36:aa4e > > > -- I use a the systemd automount functionality to mount NFS shares on demand. Use "x-systemd.automount" as one of the mount options and the problem will go away. Bonus, if the network is legitimately down, it doesn't hold up booting. `man systemd.mount` explains, http://www.freedesktop.org/software/systemd/man/systemd.mount.html . --Pete -- 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: etc-shadow
On Sun, Feb 22, 2015 at 11:23:45AM -0700, Chris Murphy wrote: > Curiously, Anaconda calls authconfig to create the key, and the resulting > shadow entry contains a 16 character salt. Whereas passwd uses an 8 > character salt. Huh, that is curious. I assume we really want to be using the 16-char salt everywhere -- bug against passwd, I guess. -- Matthew Miller Fedora Project Leader -- 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: etc-shadow
On Sun, Feb 22, 2015 at 10:36 PM, Matthew Miller wrote: > On Sun, Feb 22, 2015 at 11:23:45AM -0700, Chris Murphy wrote: >> Curiously, Anaconda calls authconfig to create the key, and the resulting >> shadow entry contains a 16 character salt. Whereas passwd uses an 8 >> character salt. > > Huh, that is curious. I assume we really want to be using the 16-char > salt everywhere -- bug against passwd, I guess. https://bugzilla.redhat.com/show_bug.cgi?id=1195110 I've also mentioned it on security@ list. -- Chris Murphy -- 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: F21 partitioning circus
On 02/22/2015 09:01 PM, Chris Murphy wrote: What Anaconda did with new UI is break with that tradition, and emphasize final results, not the nutty esoteric details of how to get there. Where it still frustrates is how it doesn't convey this worldview very well to the user. Well, I think you have to say so - My view differs ;) In my experience, anaconda is the #1 point, many people (ordinary users and power users) are complaining about when getting in contact with Fedora and is the #1 reason why they are shying away from installing Fedora (When talking to non-Fedora users, the first question very often is "Is the installer still the crap it used to be?".) To "newbies" the GUI is "cryptical" and "non-selfexplatory", while to "power-users" the GUI doesn't provide the features catering their demands and clumsy to use. That's a difficult problem to solve, the result is the user still thinks they're supposed to be able to manipulate partitions. IMO, this is a distorted view. People want to understand what the installer does and to have control over it. The current GUI does not do so and instead applies some magic which people have learnt does not do what they want. Ralf -- 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: F21 partitioning circus
On 02/23/2015 12:43 AM, Chris Murphy wrote: Look at even Android and cyanogen. Look at the reinvention of all OS's for mobile devices and how much simpler things are when constraining choices. Chromebooks are in that same category. Simple. Just works. They picked a layout and stuck with it. These OSes address the use-case of a completely non-knowledge able user to install an OS as single OS on some piece of comparatively non-complex HW. Not that this use-case would not exist with Linux, but this is very different from typical Fedora and Linux use-cases, which comprises multibooting and co-existence with other OSes, a wide variety of HW and a wide variety of configurations. Ralf -- 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: F21 partitioning circus
On Sun, 2015-02-22 at 15:01 -0700, Chris Murphy wrote: > What you're talking about might be in-scope for blivet-gui. It > definitely sounds out of scope for a GUI OS installer. > > Windows, OS X installers have maybe 2-3 total layouts between them. > And their installers are completely, totally, bullet proof. They don't > ever crash, or ask the user to create required partitions, they always > succeed in their penultimate goal which is to install a bootable OS. > And there are essentially zero user complaints about these installers. > There's nothing at all to even complain about because they don't do > anything except meet their primary requirement. Not even their > developers or testers even complain about the installer, it does one > thing successfully. While I don't find it hard to believe that Windows developers won't complain. After all, just about all Windows users do is install Windows as a new install, or over the top of a previous one, with no intention of doing anything like dual-boot. Shoe-horn it in, that's all they care about. These days, it's all single-partition, or act like it's single-partition with a hidden boot/recovery partition that the user doesn't know about. I find it harder to believe that users don't complain about the Windows installer. I've certainly seen it fuck up, and I can't be the only one. It was a gamble to see whether an install over the top could manage to keep existing data, never mind settings. And trying to get it to install to the right drive in a two-drive PC was nothing but trial and error (one drive for Windows, a second drive for video on a non-linear editing suite). I, also, am rather incredulous of how difficult it is to have the Linux installer simply do what the user tells it to do, instead of second-guessing them and denying them of what they want to do. If I select custom partition, and edit partitions myself, type of options, I expect it to have a GUI that does what I tell it to do. In the past, before the live DVD install era, I'd boot the install disc and wait for to pause on some screen, then CTRL + ALT + FUNCTION-KEY to another terminal, and fdisc my hard drive, and go back to the installer and have it use my pre-defined partitions. Even further back, I'd select the options to check partitions for faults, rather than get a nasty surprise a few months in when the drive reaches a certain amount of fullness and comes across a bad section. I don't know what's really so hard about giving us a simple GUI hard drive partitioner somewhere in the install routine. Using the command line tool is a pain (e.g. you cannot see any details about the rest of the drive while you're working on making a partition), and there are other standalone GUI partitioning tools that exist. Leave the so-called automatic smart partitioning to those people who choose the full-automatic option. -- tim@localhost ~]$ uname -rsvp Linux 3.18.7-100.fc20.i686 #1 SMP Wed Feb 11 21:16:53 UTC 2015 i686 All mail to my mailbox is automatically deleted, there is no point trying to privately email me, I will only read messages posted to the public lists. George Orwell's '1984' was supposed to be a warning against tyranny, not a set of instructions for supposedly democratic governments. -- 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
