users February 2020

users@lists.fedoraproject.org

106 participants
97 discussions

Weird dnf problem... Fixed, but weird.
by Michael H. Warfield 25 Feb '20

25 Feb '20

Hey all, Background -- I run a cluster of x86_64, armfp, and aarch64 systems and I use a shared cache on my 100TB NAS for dnf so a package (even if it's a DRPM) only has to be downloaded once. Works pretty damn well for me and I know other people have tried to use (abuse) pkgcacher and socks and straight nfs and others to do this. The dnf maintainers should take note of a need (I've browsed some of the bugzilla reports). There are a few of us out there that want to download rpm files ONCE and avoid beating the sh*t out of the mirrors but don't want to build entire mirror repositories. But that wasn't the cause of this problem. It merely tripped over it. IAC, on a couple of my Raspberry Pi systems, I started getting this weird error about an sqlite error in "Generating completion cache..." during many operations. Well, I'm pointed at an alternate cache location but something was seriously wrong here. Just plain dnf was giving me the same error, ignoring the dnf-overlay front-end. Tried all the "dnf clean" options - no joy. Tried "rpm --rebuilddb" - no joy. Finally chased down the error message in the code to this file... /usr/lib/python3.7/site-packages/dnf-plugins/generate_completion_cache.py And this line: self.cache_file = "/var/cache/dnf/packages.db" Fancy that. It apparently doesn't obey the caching options in dnf.conf or from the command line. It also doesn't get cleaned in the cleanup operations. Your only option really is to remove that db file and rerun dnf update again to rebuild it and it's cool! I've filed a bug report here: https://bugzilla.redhat.com/show_bug.cgi?id=1807238 Mostly, it can be ignored. I'm not sure WHY a bash completion database should be in a hard coded location but I'm not sure what would be better in the local system (maybe something in /var/lib) but it was frustrating to have to chase this thing down in the code when that damn database got corrupted and even worse that dbf could not clean it. In case anyone else gets bit by this... Regards, Mike -- Michael H. Warfield (AI4NB) | (o) +1 706 850-8773 | mhw(a)WittsEnd.com /\/\|=mhw=|\/\/ | (c) +1 678 463-0932 | http://www.wittsend.com/mhw/ ARIN whois: MHW9-ARIN | An optimist believes we live in the best of all PGP Key: 0xC0EB9675674627FF | possible worlds. A pessimist is sure of it!

1 0

Mate panel: false error report
by Beartooth 24 Feb '20

24 Feb '20

I got an F31 message saying Mate-panel had crashed -- and it hadn't. This is on a machine which displays both on its own monitor and on my main one. (Both are behind an HDMI KVM-switch. Each has panels on all four sides, not identical.) The box that popped up the erroneous notice asked for a report, which I started, and then wanted to upload a whole core load, at many MiB and slowly. Stopping the upload killed the popped up box. If the Mate crew really does want all that dump, please tell me how and where to send it, or what to do instead, and I will. -- Beartooth Staffwright, Not Quite Clueless Power User Remember I know little (precious little!) of where up is.

1 0

workspace switcher : minor feature request
by Beartooth 23 Feb '20

23 Feb '20

This is a minor irritation, at least afaik: I'd fix it if I could AND saw an easy way to. I live by the Fedora workspace switcher -- probably the most valuable applet I have, because it's so easy to set and forget. I have canonical spaces for my terminal (with several tabs, my other most used applet), my default browser (one that can handle long URLs, meaning Firefox or Seamonkey unless I'm missing a good third one), one other (varying) browser, my newsreader (All hail Pan!), and a space right below it for writing posts. There are also several usually free spaces scattered among those, so that I can bring up anything else conveniently, and see the canonicals out of the corner of my eye. I also upgrade every day, and reboot whenever there's a kernel change. I suppose I should close the canonical applets, but I often don't. Then when the reboot finishes, they all show up on the upper left space, and I have to rearrange them. IFF, and I do mean only if, it would be easy to make them launch where they were before, that would make the workspace switcher even more valuable. -- Beartooth Staffwright, Not Quite Clueless Power User Remember I know little (precious little!) of where up is.

3 4

how to detect hack attempts.
by home user 23 Feb '20

23 Feb '20

(F-30; Gnome; stand-alone home workstation) Sometime last year, I saw an article that talked about a tool that quickly and easily shows attempts to hack in to a computer. I think it was either in the Fedora magazine or Gnome's website. I've since made multiple attempts to find that article, but failed. I'm needing to check for hack-in attempts (something I suppose I should do quazi-periodically anyway). What is the tool/application to do that? If such a tool/application does not exist, then what is the best way for me to do that? thanks, Bill.

10 73

rkhunter after dnf update
by Dario Lesca 23 Feb '20

23 Feb '20

After any updates I must remember to run "rkhunter --propupd". Is it possible to do this automatically? There is some dnf plugins that does this? Many thanks. -- Dario Lesca (inviato dal mio Linux Fedora 31 Workstation)

1 0

nvidia and kernel 5.5
by Simon Colston 22 Feb '20

22 Feb '20

Fedora Workstation 31 I was doing a dnf upgrade (which I aborted), when I got the message: Removing dependent packages: kmod-nvidia-5.4.18-200.fc31.x86_64 x86_64 3:440.59-1.fc31 @@commandline 30 M I assume this package is dependent on kernel 5.4 and the dnf upgrade includes kernel 5.5. So, I did [simon@hedgehog ~]$ dnf list kmod-nvidia* Last metadata expiration check: 0:00:10 ago on Sat 22 Feb 2020 08:51:05 GMT. Installed Packages kmod-nvidia-5.4.18-200.fc31.x86_64.x86_64 3:440.59-1.fc31 @@commandline kmod-nvidia-5.4.19-200.fc31.x86_64.x86_64 3:440.59-1.fc31 @@commandline kmod-nvidia-5.4.20-200.fc31.x86_64.x86_64 3:440.59-1.fc31 @@commandline Available Packages kmod-nvidia.x86_64 3:440.59-1.fc31 rpmfusion-nonfree-nvidia-driver Now, the kmod-nvidia package says that it is a meta-package to ensure the nvidia module(s) for the newest kernel. So, my questions are: 1. Should I remove the installed nvidia packages and replace them with the meta-package? 2. Does the 'dnf list' imply that the nvidia packages for kernel 5.5 are not ready yet? 3. I don't want to upgrade the kernel until the nvidia package is available. What's the best way to monitor that? Thanks, Simon

5 6

OT: Pad for handwriting
by Paul Smith 22 Feb '20

22 Feb '20

Dear All, I am looking for a pad to handwrite on both Fedora and MS Windows. I am a total novice regarding that type of pads, so some help and direction will be most appreciated. The same pad is to be used also on MS Windows. Mostly, the pad is to deliver my lectures instead of using blackboard. Thanks in advance for your help, Paul

2 3

Key-Based Authentication -
by Bob Goodwin 22 Feb '20

22 Feb '20

I've been reading the thread about detecting hack attempts and I am interested in in setting up "key based authentication" as described [perhaps] in "https://docs.fedoraproject.org/en-US/Fedora/14/html/Deployment_Guide/s2-ssh…" In doing this is their danger of making an error and locking myself out of my computer, if so what to avoid? I've made some catastrophic errors in the not very distant past that required a new system re-installation and would prefer not repeating that. Suggestions, thoughts? Bob -- Bob Goodwin - Zuni, Virginia, Fedora Linux-31 XFCE

8 8

http failing on boot
by Scott van Looy 21 Feb '20

21 Feb '20

This is probably something super simple, but… My httpd is failing on boot. The server has 3x adaptors, all are static IPs and don’t require DHCP or anything. Addresses, etc are specified in the config files. All are present once boot has ended. All are managed by NetworkManager, yet I get this error in the log Feb 21 07:40:15 <server> systemd[1]: Starting The Apache HTTP Server... Feb 21 07:40:16 <server> httpd[1012]: [Fri Feb 21 07:40:16.208953 2020] [so:warn] [pid 1012:tid 140114180026688] AH01574: module ssl_module is already loaded, skipping Feb 21 07:40:16 <server> httpd[1012]: (99)Cannot assign requested address: AH00072: make_sock: could not bind to address <my ip>:80 Feb 21 07:40:16 <server> httpd[1012]: no listening sockets available, shutting down Feb 21 07:40:16 <server> httpd[1012]: AH00015: Unable to open logs Feb 21 07:40:16 <server> systemd[1]: httpd.service: Main process exited, code=exited, status=1/FAILURE Feb 21 07:40:16 <server> systemd[1]: httpd.service: Failed with result 'exit-code'. Feb 21 07:40:16 <server> systemd[1]: Failed to start The Apache HTTP Server. It’s like httpd is starting too early. Looking at the service info I can see it’s supposed to start after network.target. Looking at its critical chain using systemd-analyze it seems to be doing this correctly. I’m kinda at a loss. Does anyone have any suggestions as to what I can try? Scott

3 2

systemd-tty-ask-password-agent?
by Tom Horsley 20 Feb '20

20 Feb '20

An infection seems to be spreading in systemd. First I saw dhcpd taking forever to shut down: https://bugzilla.redhat.com/show_bug.cgi?id=1768604 Now I just saw the exact same thing with the apache httpd service. I found the systemctl --no-ask-password option, so I tried it, now it no longer forks and execs the password-agent process, it merely sits like a lump timing out for the same amount of time. Anyone have any idea what leads systemd to decide it needs to wait for some mysterious something when stopping some services? And why is the number of these mysterious services growing? (Does this have something to do with me not using NetworkManager?)

4 9

← Newer
1
2
3
4
5
6
7
8
9
10
Older →