Bug#504000: Works for me
> It does. This may be related to a known upstream problem with some > motherboards. Please try commenting out the rtcfile directive in > /etc/chrony/chrony.conf. After commenting out rtcfile upgrading the package again works, so that might at least be a work around. Still I think that this regression should be possible to fix, because 1.21z-5 is not hit. Helmut -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
> I had problems with /dev/rtc before, sometimes related to HPET which in > combination with chrony even froze my system. This kernel bug has been fixed > recently. Also I had problems when using the wrong module. Are you sure you > use > the right one? Does hwclock work for you? I don't really know about rtc stuff, but I found out that /dev/rtc is driven by the module rtc_cmos on my system, because the usage count of that module increases on opening the device. hwclock works. I tried reading and writing. Helmut -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
Helmut writes: > Exactly: "or is hung". An hung is a bug. Yes, of course it is. However, it is important to know whether Chronyd is failing to start, starting and then crashing, or getting hung up. -- John Hasler [EMAIL PROTECTED] Elmwood, WI USA -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
Helmut writes: > I also tried to strace chronyd directly. It hangs in a read from > filedescriptor 6 which happens to be /dev/rtc according to lsof, so > indeed chronyd *is* hung. Hope this helps. It does. This may be related to a known upstream problem with some motherboards. Please try commenting out the rtcfile directive in /etc/chrony/chrony.conf. -- John Hasler [EMAIL PROTECTED] Elmwood, WI USA -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
On Fri, Oct 31, 2008 at 10:22:40PM +0100, Helmut Grohne wrote: > I also tried to strace chronyd directly. It hangs in a read from > filedescriptor 6 which happens to be /dev/rtc according to lsof, so > indeed chronyd *is* hung. Hope this helps. I had problems with /dev/rtc before, sometimes related to HPET which in combination with chrony even froze my system. This kernel bug has been fixed recently. Also I had problems when using the wrong module. Are you sure you use the right one? Does hwclock work for you? Michael -- Michael Meskes Michael at Fam-Meskes dot De, Michael at Meskes dot (De|Com|Net|Org) Michael at BorussiaFan dot De, Meskes at (Debian|Postgresql) dot Org ICQ: 179140304, AIM/Yahoo: michaelmeskes, Jabber: [EMAIL PROTECTED] Go VfL Borussia! Go SF 49ers! Use Debian GNU/Linux! Use PostgreSQL! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
Hi John, > > When I manually start chronyc and then type a command (sourcestats for > > instance) it keeps waiting forever. > This almost certainly means that chronyd is not running (or is hung). Exactly: "or is hung". An hung is a bug. Helmut -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
Hello John, > > I had to downgrade, because chrony also broke suspend2disk[1]... > Is this a different bug? If so it should be reported seperately. I explained why this affected ... > > ...so I could as well have tagged it critical because it affects other > > applications as well. > I can find no package named anything like "suspend2disk". As I have no > laptop I am not familiar with the various hibernation and suspension > schemes but the Chrony package includes this script: ... hibernate in the footnote. The term suspend2disk was used as a functional term. It is just ... > > I use hibernate and restart chrony on wakeup to synchronize with > > new servers. As chrony does not terminate... > What do you mean by "does not terminate"? that chronyc (sorry for being unclear) does not terminate, and that way the init script invoking ppp/ip-up.d/chrony does not terminate and that way the hibernate hook does not terminate and that way hibernate does not terminate and that way refuses to be run in parallel therefore breaking it. Just forget about this and let us fix the bug. Helmut -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
Hallo Michael, > Have you ever tried using the default config? Yes, no change. I also tried to strace chronyd directly. It hangs in a read from filedescriptor 6 which happens to be /dev/rtc according to lsof, so indeed chronyd *is* hung. Hope this helps. Helmut -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
On Fri, Oct 31, 2008 at 09:45:45AM -0500, John Hasler wrote: > > ...so I could as well have tagged it critical because it affects other > > applications as well. > > I can find no package named anything like "suspend2disk". As I have no > laptop I am not familiar with the various hibernation and suspension > schemes but the Chrony package includes this script: > ... I do have one where I use chrony. As mentioned before it works flawlessly. > Are you saying that the script is not working, that suspend2disk needs > something else, or that Chrony fails to stop with "/etc/init.d/chrony > stop"? if it is the second it is not a bug in Chrony. I think what he meant was that this bug we are talking about causes the whole hibernation process to stop, which is quite obvious when /etc/ppp/ip-up.d/chrony hangs for whatever reason. Michael -- Michael Meskes Michael at Fam-Meskes dot De, Michael at Meskes dot (De|Com|Net|Org) Michael at BorussiaFan dot De, Meskes at (Debian|Postgresql) dot Org ICQ: 179140304, AIM/Yahoo: michaelmeskes, Jabber: [EMAIL PROTECTED] Go VfL Borussia! Go SF 49ers! Use Debian GNU/Linux! Use PostgreSQL! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
On Fri, Oct 31, 2008 at 02:49:04PM +0100, Helmut Grohne wrote: > Thanks for you effort. If you have any ideas for debugging, please tell > me. Have you ever tried using the default config? Michael -- Michael Meskes Michael at Fam-Meskes dot De, Michael at Meskes dot (De|Com|Net|Org) Michael at BorussiaFan dot De, Meskes at (Debian|Postgresql) dot Org ICQ: 179140304, AIM/Yahoo: michaelmeskes, Jabber: [EMAIL PROTECTED] Go VfL Borussia! Go SF 49ers! Use Debian GNU/Linux! Use PostgreSQL! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
Helmut writes: > When I manually start chronyc and then type a command (sourcestats for > instance) it keeps waiting forever. This almost certainly means that chronyd is not running (or is hung). -- John Hasler [EMAIL PROTECTED] Elmwood, WI USA -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
Helmut writes: > I had to downgrade, because chrony also broke suspend2disk[1]... Is this a different bug? If so it should be reported seperately. > ...so I could as well have tagged it critical because it affects other > applications as well. I can find no package named anything like "suspend2disk". As I have no laptop I am not familiar with the various hibernation and suspension schemes but the Chrony package includes this script: #!/bin/sh # Placed in /etc/apm/event.d by the chrony package at the instruction of # the apmd maintainer. If you don't have apm and don't intend to install # apmd you may remove it. It needs to run after 00hwclock but before any # other scripts. [ -x /usr/sbin/chronyd ] || exit 0 if [ "$1" = suspend ]; then /etc/init.d/chrony stop elif [ "$1" = standby ]; then /etc/init.d/chrony stop elif [ "$1" = resume ]; then /etc/init.d/chrony start fi Are you saying that the script is not working, that suspend2disk needs something else, or that Chrony fails to stop with "/etc/init.d/chrony stop"? if it is the second it is not a bug in Chrony. > I use hibernate and restart chrony on wakeup to synchronize with > new servers. As chrony does not terminate... What do you mean by "does not terminate"? -- John Hasler [EMAIL PROTECTED] Elmwood, WI USA -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
Hallo Michael, > Sorry, I don't buy into this argument, but it's nitpicking anyway. The bug has > to be fixed. ACK > > It is not empty. And if it was empty, downgrading would not solve the > > issue. > > No, that's not necessarily true. There may be a problem in the shell script > for > example or in the package config. After all the password is automatically read > from the configs. Ok. Let me mention another observation I should have mentioned before: When I manually start chronyc and then type a command (sourcestats for instance) it keeps waiting forever. I also tried to strace chronyc and I haven't seen syscalls receiving data. > Is it possible for you to upgrade, test and downgrade again? Tried several times (at least accidentally ;-). > Please keep in mind that I'm not the maintainer but just looking into some > critical bugs in software that I use. Thanks for you effort. If you have any ideas for debugging, please tell me. Helmut -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
On Fri, Oct 31, 2008 at 01:33:48PM +0100, Helmut Grohne wrote: > If you want to get an answer I urge you to also contact the submitter > (me) by either by ccing me directly or by ccing [EMAIL PROTECTED] Argh, sorry, just forgot it. > I had to downgrade, because chrony also broke suspend2disk[1], so I > could as well have tagged it critical because it affects other Sorry, I don't buy into this argument, but it's nitpicking anyway. The bug has to be fixed. > applications as well. Still I have to admit that this version also runs > on my sid/amd64 machine and works well there. If it only hits some system I'd say it's not RC, but anyway. > It is not empty. And if it was empty, downgrading would not solve the > issue. No, that's not necessarily true. There may be a problem in the shell script for example or in the package config. After all the password is automatically read from the configs. Is it possible for you to upgrade, test and downgrade again? Please keep in mind that I'm not the maintainer but just looking into some critical bugs in software that I use. Michael -- Michael Meskes Michael at Fam-Meskes dot De, Michael at Meskes dot (De|Com|Net|Org) Michael at BorussiaFan dot De, Meskes at (Debian|Postgresql) dot Org ICQ: 179140304, AIM/Yahoo: michaelmeskes, Jabber: [EMAIL PROTECTED] Go VfL Borussia! Go SF 49ers! Use Debian GNU/Linux! Use PostgreSQL! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
If you want to get an answer I urge you to also contact the submitter (me) by either by ccing me directly or by ccing [EMAIL PROTECTED] Otherwise it is very unlikely that I eventually catch your comment. See #351856. > Given that the whole setup works like a charm for me I think the severity is > too high, thus I downgraded it. From looking at /etc/ppp/ip-up.d/chrony I'd I had to downgrade, because chrony also broke suspend2disk[1], so I could as well have tagged it critical because it affects other applications as well. Still I have to admit that this version also runs on my sid/amd64 machine and works well there. > guess that $PASSWORD is empty for you. Could you please verify? It is not empty. And if it was empty, downgrading would not solve the issue. Any other suggestions? Helmut [1] I use hibernate and restart chrony on wakeup to synchronize with new servers. As chrony does not terminate, hibernate doesn't either and as long as hibernate runs, it will not hibernate again. So hibernate got borked, too. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#504000: Works for me
severity 504000 important thanks Given that the whole setup works like a charm for me I think the severity is too high, thus I downgraded it. From looking at /etc/ppp/ip-up.d/chrony I'd guess that $PASSWORD is empty for you. Could you please verify? Michael -- Michael Meskes Michael at Fam-Meskes dot De, Michael at Meskes dot (De|Com|Net|Org) Michael at BorussiaFan dot De, Meskes at (Debian|Postgresql) dot Org ICQ: 179140304, AIM/Yahoo: michaelmeskes, Jabber: [EMAIL PROTECTED] Go VfL Borussia! Go SF 49ers! Use Debian GNU/Linux! Use PostgreSQL! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
