random process crashes in virtualbox guests (clocksource problem?)
hey, i run debian 11 on both host and guest and experience random process crashes in virtual box guests under load. i.e. just reading a lot of files from any disk on the host system causes the guests to randomly crash a process... e.g. if i move a large file # mv -nv or if i do just # md5sum in fact, it doesn't really matter what i do, as long as it creates significant i/o traffic and is lasting longer than a couple of seconds. in any case, the guest starts to complain with: Jul 01 20:45:20 vmguest kernel: clocksource: timekeeping watchdog on CPU0: kvm-clock wd-wd read-back delay of 816732ns Jul 01 20:45:20 vmguest kernel: clocksource: wd-tsc-wd read-back delay of 2540783ns, clock-skew test skipped! and, what is worse, processes start to randomly crash: Jun 30 20:17:01 vmguest kernel: traps: sh[250946] general protection fault ip:7fb0c341708e sp:7ffec3154378 error:0 in libc-2.31.so[7fb0c33a6000+14b000] Jul 01 00:00:02 vmguest kernel: traps: hostname[253617] general protection fault ip:7f905f2b24a6 sp:7fff44a30e30 error:0 in libc-2.31.so[7f905f299000+14b000] Jul 01 00:53:01 vmguest kernel: traps: wget[254290] general protection fault ip:7f934bc23fda sp:7ffd716954d0 error:0 in libtasn1.so.6.6.0[7f934bc1a000+c000] does anyone have any idea what the problem is? or where to turn to, do find a solution? the only thing i found was an unresolved issue in the debian forum: https://forums.debian.net/viewtopic.php?t=151851 any help would be much appreciated! greetings...
Re: Debian 11: synaptic fails to fetch fetchmail
On Sun, 3 Jul 2022, to...@tuxteam.de wrote: My hunch is that your synaptic is looking at a stale package database. Do an "apt-get update" or whatever you have to do to synaptic to achieve the same effect, perhaps the problem goes away. Yes, I did "apt-get update" and the synaptic problem went away. Thanks. Roger
Re: Upgrade issue with Debian 9 -> 10
On 7/3/22 02:31, Miroslav Skoric wrote: Hi all, Yesterday I attempted to upgrade Compaq Presario CQ56 laptop to buster. I followed instructions in 'Chapter 4. Upgrades from Debian 9 (stretch)', so all went well with a minimal upgrade (apt-get upgrade). When it finished, I went to the main part of the upgrade (apt full-upgrade). It ran well until some 40-45% and then started complaining about lack of disk space. (apt -o APT::Get::Trivial-Only=true full-upgrade did not say I shall get into any trouble.) So, at one point the full upgrade just exited. I tried to uninstall some old stuff but it was not possible. df -h showed that / and /usr were almost 100% used. Shutdown & reboot seemed going normally, although including few [FAILED] warnings mostly with firewall failed to start and like. Majority went [OK] until the point where it was about to perform fsck on mounted volumes where it looks as an endless process occasionally repeating this line: [nnn.nn] perf: interrupt took too long ( > ), lowering kernel.perf_event_max_sample_rate to n where 'n' are numbers. Ctrl-Alt-F2 brings tty2 from where I can log in, then sudo etc. df -h shows that filesystem /dev/mapper/localhost-root (mounted on /) is 99% used, and /dev/mapper/localhost-usr (mounted on /usr) is 100% used. As it is (an encrypted) LVM, where /dev/mapper/localhost-home (mounted on /home) is only 21% used, I suppose that it shall be possible to resize partitions i.e. logical volumes so that some space of /home to be assigned to / and /usr It seems that resize2fs, lvextend, and some related commands are available in tty2, but I am unsure about the proper order & syntax of those commands. Also, what about the ongoing fsck process in tty1? Any suggestion? The KISS approach is to check in your system configuration files to a version control system, back up your data, take an image of the OS drive, remove the OS drive, insert a blank OS drive, do a fresh install, check out the old system configuration files to a side directory, configure the new OS instance, restore your data, and validate everything. David
Re: Upgrade issue with Debian 9 -> 10
On 7/3/22 1:17 PM, Andrew M.A. Cater wrote: Ctrl-Alt-F2 brings tty2 from where I can log in, then sudo etc. df -h shows that filesystem /dev/mapper/localhost-root (mounted on /) is 99% used, and /dev/mapper/localhost-usr (mounted on /usr) is 100% used. Apt tends to store files in /var - it's possible that /var is also full. Possibly. /var was always around 49-50% used here, but I knew from some earlier upgrade that it might be too small to store new packages for upgrading to buster. And because of that I added a thumb drive as a temporary /var/archives, and it served the purpose. If you repeat an apt-get update - do you have errors about needing to rerun a configure step? Haven't tried that, but something else already helped: While it was idling with fsck in tty1, I went to tty2 and entered: apt --fix-broken install ... and it did/resumed full upgrade. (Interestingly, this time it did not complain about no space in / and /usr.) When it finished, I tested startx and it brought GUI. Not sure now but I think that I then rebooted and it went it into GUI as expected. So far - so good. Few red [FAILED] warnings during CLI phase related to not starting UFV, Shorewall, and minissdpd services, so I need to check for that. A subsequent apt --fix-broken install (or some other command) only complained about some initrd issue with kernel image 4.19.0-20-686 so I removed that image and stayed with 4.9.0-19-686. After that, apt autoremove freed some 500MB of old stretch packages so now / is about 97% used, while /usr is still 100% used. In thi situation, I might be tempted to save off any data in /home and any options in /etc/ to configure mail and things like that and do a reinstall with Debian 11 as a quick fix but that's a destructive option. Will see whether it will work without such a destructive option :-) In fact, that laptop started running Linux as squeeze some ten years ago, and I gradually upgraded it to wheezy, jessie, ... It makes me wonder if I gave it few more years of life with buster, after stretch went EOL yesterday. When you installed, did you manually specify sizes for filesystems or did you say "install in one encrypted LVM"? I cannot remember because I made it with squeeze somewhere in 2013 or so. What I do recall is that at some upgrade point I had similar space issues, when resize2fs and/or lvextend solved the problem within the existing LVM. (I had 'borrowed' some space where I had a surplus, and added where needed. Probably I will need to learn it again.) If you did that then, effectively, /home and so on are auto-sized and LVM is keeping track of free space. Deleting unwanted files is the only way to reclaim space and then, perhaps resize. Well, for sure I missed to uninstall some software that I rarely used in stretch, and if I did so I might have not got into trouble. Now will take more care with buster. Good luck with it all - with every good wish, as ever, Andy Cater Thanks. Misko
Re: Problem with csh
Deleted ~/.csh. Problem solved. On 07/02/2022 05:23 PM, Will Mengarini wrote: On Sat, Jul 02, 2022 at 01:18:08PM -0400, Stephen P. Molnar wrote: comp@AbNormal:~$ csh Bad : modifier in $ '/'. On Sat, Jul 02, 2022 at 03:31:12PM -0400, Bijan Soleymani wrote: Quick Google search shows it is an issue with the syntax of defining environment variables: https://stackoverflow.com/questions/40968061/bad-modifier-in * Greg Wooledge [22-07/02=Sa 15:41 -0400]: That still requires some command to have been executed. Since Stephen didn't even run a command yet, that means he has to have created a bogus dot file (e.g. ~/.cshrc) containing the invalid csh command. There could be an error in a system-wide init file. He says this is a fresh Bullseye, but not that he's the sysadmin who set it up, and he's in an academic environment. On a Manjaro system that I don't administer I saw the line set -r autologout 86400 which should be set -r autologout = 86400 and after the sysadmin fixed it, a Manjaro update broke it again. Because csh gets little use nowadays, it's possible there's something wrong in Debian's init files too (not the set -r error though, because that gives a different error message), though I assume that whatever is wrong would need to be on a code path not followed for most testing. That seems plausible, since the behavior of startup files could depend on hardware configuration. -- Stephen P. Molnar, Ph.D. Molecular Modeling 614.312.7528 (c) Skype: smolnar1
Re: Debian 11: synaptic fails to fetch fetchmail
On Sun, 3 Jul 2022 at 21:11, Roger Price wrote: > I would like to install fetchmail on Debian 11, but synaptic gives me the > following error message: > > W: Failed to fetch > http://deb.debian.org/debian/pool/main/f/fetchmail/fetchmail_6.4.16-4_amd64.deb >404 Not Found [IP: 199.232.178.132 80] > > Is this temporary or do I need to look elsewhere for fetchmail .deb? Roger # apt update # apt-get --print-uris --no-download download fetchmail 'http://deb.debian.org/debian/pool/main/f/fetchmail/fetchmail_6.4.16-4%2bdeb11u1_amd64.deb' fetchmail_6.4.16-4+deb11u1_amd64.deb 401968 SHA256:0ce0a934de679625b14254dee468e9b35fe3979e1ff6272555acbfed88a1bee7
Re: Debian 11: synaptic fails to fetch fetchmail
On Sun, Jul 03, 2022 at 01:11:04PM +0200, Roger Price wrote: > I would like to install fetchmail on Debian 11, but synaptic gives me the > following error message: > > W: Failed to fetch > http://deb.debian.org/debian/pool/main/f/fetchmail/fetchmail_6.4.16-4_amd64.deb > 404 Not Found [IP: 199.232.178.132 80] > > Is this temporary or do I need to look elsewhere for fetchmail .deb? Roger Looking around here http://deb.debian.org/debian/pool/main/f/fetchmail/ reveals that the package name is fetchmail_6.4.16-4+deb11u1_amd64.deb these days. My hunch is that your synaptic is looking at a stale package database. Do an "apt-get update" or whatever you have to do to synaptic to achieve the same effect, perhaps the problem goes away. Cheers -- t signature.asc Description: PGP signature
Re: Upgrade issue with Debian 9 -> 10
On Sun, Jul 03, 2022 at 11:31:40AM +0200, Miroslav Skoric wrote: > Hi all, > > Yesterday I attempted to upgrade Compaq Presario CQ56 laptop to buster. I > followed instructions in 'Chapter 4. Upgrades from Debian 9 (stretch)', so > all went well with a minimal upgrade (apt-get upgrade). When it finished, I > went to the main part of the upgrade (apt full-upgrade). It ran well until > some 40-45% and then started complaining about lack of disk space. > > (apt -o APT::Get::Trivial-Only=true full-upgrade did not say I shall get > into any trouble.) > > So, at one point the full upgrade just exited. I tried to uninstall some old > stuff but it was not possible. df -h showed that / and /usr were almost 100% > used. > > Shutdown & reboot seemed going normally, although including few [FAILED] > warnings mostly with firewall failed to start and like. Majority went [OK] > until the point where it was about to perform fsck on mounted volumes where > it looks as an endless process occasionally repeating this line: > > [nnn.nn] perf: interrupt took too long ( > ), lowering > kernel.perf_event_max_sample_rate to n > > where 'n' are numbers. > > Ctrl-Alt-F2 brings tty2 from where I can log in, then sudo etc. df -h shows > that filesystem /dev/mapper/localhost-root (mounted on /) is 99% used, and > /dev/mapper/localhost-usr (mounted on /usr) is 100% used. > Apt tends to store files in /var - it's possible that /var is also full. If you repeat an apt-get update - do you have errors about needing to rerun a configure step? In thi situation, I might be tempted to save off any data in /home and any options in /etc/ to configure mail and things like that and do a reinstall with Debian 11 as a quick fix but that's a destructive option. apt-get clean may clear out some downloaded packages and provide some space. > As it is (an encrypted) LVM, where /dev/mapper/localhost-home (mounted on > /home) is only 21% used, I suppose that it shall be possible to resize > partitions i.e. logical volumes so that some space of /home to be assigned > to / and /usr > When you installed, did you manually specify sizes for filesystems or did you say "install in one encrypted LVM"? If you did that then, effectively, /home and so on are auto-sized and LVM is keeping track of free space. Deleting unwanted files is the only way to reclaim space and then, perhaps resize. There's a reason that I install into one filesystem if I can - manual sizing and partitioning rarely works unless you have a specific use. On one machine here I have a 7TB /srv partition deliberately because it's full of data that I want to serve out via a webserver - in any other machine, I'd probably have said use the whole 8TB filesystem and auto partition. > It seems that resize2fs, lvextend, and some related commands are available > in tty2, but I am unsure about the proper order & syntax of those commands. > Also, what about the ongoing fsck process in tty1? Any suggestion? > Good luck with it all - with every good wish, as ever, Andy Cater > Misko >
Debian 11: synaptic fails to fetch fetchmail
I would like to install fetchmail on Debian 11, but synaptic gives me the following error message: W: Failed to fetch http://deb.debian.org/debian/pool/main/f/fetchmail/fetchmail_6.4.16-4_amd64.deb 404 Not Found [IP: 199.232.178.132 80] Is this temporary or do I need to look elsewhere for fetchmail .deb? Roger
Upgrade issue with Debian 9 -> 10
Hi all, Yesterday I attempted to upgrade Compaq Presario CQ56 laptop to buster. I followed instructions in 'Chapter 4. Upgrades from Debian 9 (stretch)', so all went well with a minimal upgrade (apt-get upgrade). When it finished, I went to the main part of the upgrade (apt full-upgrade). It ran well until some 40-45% and then started complaining about lack of disk space. (apt -o APT::Get::Trivial-Only=true full-upgrade did not say I shall get into any trouble.) So, at one point the full upgrade just exited. I tried to uninstall some old stuff but it was not possible. df -h showed that / and /usr were almost 100% used. Shutdown & reboot seemed going normally, although including few [FAILED] warnings mostly with firewall failed to start and like. Majority went [OK] until the point where it was about to perform fsck on mounted volumes where it looks as an endless process occasionally repeating this line: [nnn.nn] perf: interrupt took too long ( > ), lowering kernel.perf_event_max_sample_rate to n where 'n' are numbers. Ctrl-Alt-F2 brings tty2 from where I can log in, then sudo etc. df -h shows that filesystem /dev/mapper/localhost-root (mounted on /) is 99% used, and /dev/mapper/localhost-usr (mounted on /usr) is 100% used. As it is (an encrypted) LVM, where /dev/mapper/localhost-home (mounted on /home) is only 21% used, I suppose that it shall be possible to resize partitions i.e. logical volumes so that some space of /home to be assigned to / and /usr It seems that resize2fs, lvextend, and some related commands are available in tty2, but I am unsure about the proper order & syntax of those commands. Also, what about the ongoing fsck process in tty1? Any suggestion? Misko