Bug#897123: linux-image-4.15.0-3-amd64: Diskless workstation regularly stalls with 4.15 kernel
A precision I have forgotten. I have filled a bug against kernel as nfs's packages haven't been upgraded when this bug was triggered. Only one workaround : use nolock mount option. Best regards, JKB
Bug#897123: linux-image-4.15.0-3-amd64: Diskless workstation regularly stalls with 4.15 kernel
Package: src:linux Version: 4.15.17-1 Severity: important Dear Maintainer, I use some diskless workstations for a long time without any specific trouble. Since I have installed 4.15 kernel from debian testing, my main workstation regularly stalls. In syslog, I have a lot of : Apr 28 18:51:53 hilbert kernel: [287335.545559] xs_tcp_setup_socket: connect returned unhandled error -107 Apr 28 18:51:55 hilbert kernel: [287337.057907] lockd: server 192.168.10.128 not responding, still trying Apr 28 18:52:07 hilbert kernel: [287349.418038] lockd: server 192.168.10.128 OK Apr 28 18:52:16 hilbert kernel: [287358.042284] lockd: server 192.168.10.128 OK Apr 28 18:52:35 hilbert kernel: [287377.715081] lockd: server 192.168.10.128 not responding, still trying Apr 28 18:53:10 hilbert kernel: [287412.071680] lockd: server 192.168.10.128 OK Apr 28 18:53:10 hilbert kernel: [287412.072447] lockd: server 192.168.10.128 not responding, still trying Apr 28 18:53:10 hilbert kernel: [287412.323525] lockd: server 192.168.10.128 not responding, still trying Apr 28 18:53:33 hilbert kernel: [287435.567724] lockd: server 192.168.10.128 not responding, still trying Apr 28 18:53:37 hilbert kernel: [287439.864192] lockd: server 192.168.10.128 not responding, still trying Apr 28 18:54:03 hilbert kernel: [287465.336821] lockd: server 192.168.10.128 OK Apr 28 18:54:09 hilbert kernel: [287471.648970] lockd: server 192.168.10.128 OK nfsstats returns : hilbert:[~] > nfsstat -m / from 192.168.10.128:/srv/hilbert Flags: rw,relatime,vers=3,rsize=65536,wsize=65536,namlen=255,hard,nolock,proto=tcp,port=2049,timeo=7,retrans=10,sec=sys,local_lock=all,addr=192.168.10.128 /home from 192.168.10.128:/home Flags: rw,relatime,vers=3,rsize=65536,wsize=65536,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=192.168.10.128,mountvers=3,mountport=1020,mountproto=tcp,local_lock=none,addr=192.168.10.128 Of course, server's configuration has not changed. This server runs NetBSD 8 and I haven't seen any NFS malfunction with kernel 4.14. Of course, all other clients running Linux (arm kernel 4.9, x86 kernel 4.14 or FreeBSD 11.1) run as expected without any trouble. Best regards, JKB -- Package-specific info: ** Version: Linux version 4.15.0-3-amd64 (debian-kernel@lists.debian.org) (gcc version 7.3.0 (Debian 7.3.0-16)) #1 SMP Debian 4.15.17-1 (2018-04-19) ** Command line: BOOT_IMAGE=pxelinux.cfg/vmlinuz-4.15.0-3-amd64-hilbert root=/dev/nfs initrd=pxelinux.cfg/initrd.img-4.15.0-3-amd64-hilbert nfsroot=192.168.10.128:/srv/hilbert ip=dhcp rw ** Tainted: O (4096) * Out-of-tree module has been loaded. ** Kernel log: [285673.025753] xs_tcp_setup_socket: connect returned unhandled error -107 [285857.121396] lockd: server 192.168.10.128 not responding, still trying [285857.373345] lockd: server 192.168.10.128 not responding, still trying [285858.641114] lockd: server 192.168.10.128 not responding, still trying [285866.965630] lockd: server 192.168.10.128 not responding, still trying [285868.225678] lockd: server 192.168.10.128 not responding, still trying [285938.011610] lockd: server 192.168.10.128 OK [285939.019589] lockd: server 192.168.10.128 OK [285947.627997] lockd: server 192.168.10.128 OK [285949.139779] lockd: server 192.168.10.128 OK [285956.528084] lockd: server 192.168.10.128 OK [286860.226436] xs_tcp_setup_socket: connect returned unhandled error -107 [286958.020204] lockd: server 192.168.10.128 not responding, still trying [286961.808706] lockd: server 192.168.10.128 not responding, still trying [286971.148954] lockd: server 192.168.10.128 OK [286994.873491] lockd: server 192.168.10.128 not responding, still trying [287048.118707] lockd: server 192.168.10.128 not responding, still trying [287082.463590] lockd: server 192.168.10.128 not responding, still trying [287116.776538] lockd: server 192.168.10.128 OK [287117.784429] lockd: server 192.168.10.128 OK [287118.820500] lockd: server 192.168.10.128 OK [287126.152638] lockd: server 192.168.10.128 OK [287250.911579] lockd: server 192.168.10.128 not responding, still trying [287251.919620] lockd: server 192.168.10.128 not responding, still trying [287264.540008] lockd: server 192.168.10.128 OK [287264.540141] xs_tcp_setup_socket: connect returned unhandled error -107 [287275.144907] lockd: server 192.168.10.128 OK [287335.545548] lockd: server 192.168.10.128 not responding, still trying [287335.545559] xs_tcp_setup_socket: connect returned unhandled error -107 [287337.057907] lockd: server 192.168.10.128 not responding, still trying [287349.418038] lockd: server 192.168.10.128 OK [287358.042284] lockd: server 192.168.10.128 OK [287377.715081] lockd: server 192.168.10.128 not responding, still trying [287408.271051] lockd: server 192.168.10.128 not responding, still trying [287412.071680] lockd: server 192.168.10.128 OK [287412.072447] lockd: server 192.168.10.128 not responding, still trying [287412.323525] lockd: server 192.168.10.128 not responding, still trying
Re: sparc / Problems with 3.2-kernel
Artyom Tarasenko a écrit : On Sun, Jun 2, 2013 at 11:38 PM, BERTRAND Joël joel.bertr...@systella.fr wrote: Andreas Barth a écrit : Hi, today I tried to resurrect our buildd on schroeder (which is from architecture sparc). While trying to do so, I had a couple of strange behaviours like vi freezing after 20-30 seconds, or I couldn't about tail with ctrl+c or suspend with ctrl+z. This didn't change after a reboot. Machine was running our official 3.2-kernel. Was the machine available via the network at the time of freeze? After Martin rebooted the machine into oldstables 2.6.32 kernel, things are working as they should, and the buildd is up again. Question: Is this a known issue? How can we get a fix for that? (I would assume DSA will consider it a non-option to not be able to upgrade to our default kernels.) I have seen same freezes on Blade 2000 (2 * US III+/900 Cu), on U5 (US IIi/440), but not on a server that run with four US IIIi and not on a T1000 (sun4v/US T1). I have tested 3.2 debian kernels and official ones and both hang. I suppose there is another strange bug with special MMU. 2.6.32 is stable but on sparc64, 3.2 is not. I have not tested newer kernels. Bertrand, Andreas, can you please enable the magic SysRq key and try to see what is the kernel busy with when it hangs? When freeze occurs, workstation was totaly blocked. SysRq didn't work anymore. I have reinstalled my Blade 2000 with NetBSD but I can try to obtain more information in a couple of weeks. My U5 runs with jessie (and a 2.6.32 kernel). Regards, JKB -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/51acd1b3.1020...@systella.fr
Re: sparc / Problems with 3.2-kernel
Andreas Barth a écrit : Hi, today I tried to resurrect our buildd on schroeder (which is from architecture sparc). While trying to do so, I had a couple of strange behaviours like vi freezing after 20-30 seconds, or I couldn't about tail with ctrl+c or suspend with ctrl+z. This didn't change after a reboot. Machine was running our official 3.2-kernel. After Martin rebooted the machine into oldstables 2.6.32 kernel, things are working as they should, and the buildd is up again. Question: Is this a known issue? How can we get a fix for that? (I would assume DSA will consider it a non-option to not be able to upgrade to our default kernels.) Hello, I have seen same freezes on Blade 2000 (2 * US III+/900 Cu), on U5 (US IIi/440), but not on a server that run with four US IIIi and not on a T1000 (sun4v/US T1). I have tested 3.2 debian kernels and official ones and both hang. I suppose there is another strange bug with special MMU. 2.6.32 is stable but on sparc64, 3.2 is not. I have not tested newer kernels. Regards, JKB -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/51abbb4f.5010...@systella.fr
Bug#400372: dpkg randomly craches on Sparc32 running HyperSPARC processor
Jurij Smakov a écrit : Hi Joel, Hello Jurij, Sorry, I cannot reproduce the behavior you are describing. I have a SS20 box, running sid with Debian's linux-image-2.6.18-3-sparc32 (version 2.6.18-6) kernel: I don't try with debian package, only with official linux kernel. I have tested the 2.6.19 but sunlance doesn't work for me in a SS20 that runs with two sunlance (eth0/1)and one hme (eth2) interfaces... I don't know the difference between the debian and official kernels. debian:~# uname -a Linux debian 2.6.18-3-sparc32 #1 Fri Nov 24 16:10:16 GMT 2006 sparc GNU/Linux debian:~# cat /proc/version Linux version 2.6.18-3-sparc32 (Debian 2.6.18-6) ([EMAIL PROTECTED]) (gcc version 4.1.2 20061115 (prerelease) (Debian 4.1.1-20)) #1 Fri Nov 24 16:10:16 GMT 2006 debian:~# cat /proc/cpuinfo cpu : ROSS HyperSparc RT625 or RT626 fpu : ROSS HyperSparc combined IU/FPU promlib : Version 3 Revision 2 prom: 2.25 type: sun4m ncpus probed: 2 ncpus active: 1 CPU0Bogo: 133.12 CPU0ClkTck : 13300 MMU type: ROSS HyperSparc contexts: 4096 nocache total : 2252800 nocache used: 492800 To stress-test dpkg I've just ran 'apt-get install kde' on an unstable system. That a pretty big update: Is your kernel SMP ? Do you use HIGHMEM ? [..] 0 upgraded, 400 newly installed, 0 to remove and 2 not upgraded. Need to get 242MB of archives. After unpacking 678MB of additional disk space will be used. and it completed without any problems. I also do not recall having any problems with dpkg recently. What OBP version do you have in these machines? I have a 2.25 from Sun in one, a 2.25R from ROSS in another one and a 2.25W (?) from an HyperSTATION in the third one. Tested modules are single and dual RT-626 with a VSIMM (4 and 8MB) and 448 MB. I think it is not an hardware trouble because all configurations I have tried return exactly the same error. Hardware of the main station I use for tests are validated with Solaris9 and without any trouble during several days (but I cannot use three or four CPU with Solaris9 without having Watchdog reset!, all combinaisons with more than two CPU are not stable. Same results on the three stations. If I have time, I shall try to install a Solaris 2.7...). ISTR that to run latest HyperSPARC CPUs from Ross you need either 2.25 from Sun, or 2.25R from Ross. If you are running anything lower, I suggest you try to upgrade your firmware, see http://www.sunshack.org/data/bootroms.html for details. Maybe this trouble is related to the trouble I see with two SuperSPARC's (Oops in readv_pipe). I don't know... Regards, JKB
Bug#363344: initramfs-tools and HyperSPARC processor
Package: initramfs-tools Version: 0.59b Severity: grave Arch: sparc It is impossible to build a ramfs image on a HyperSPARC workstation. I have try to install etch/sparc on a SS20 that runs with four HyperSPARC processors. When it boots, system hangs when it tries to load the ramfs image. If I replace HyperSPARC by SuperSPARC-II, I can restart the workstation without any trouble. I have tested some configurations and some differents sparc32 CPU's. The only combinaison that fails is ramfs created by initramfs-tools with HyperSPARC. Regards, JKB -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#363344: initramfs-tools and HyperSPARC processor
maximilian attems a écrit : severity 363344 important tags 363344 moreinfo thanks cher bertrand, On Tue, Apr 18, 2006 at 06:04:08PM +0200, BERTRAND Joël wrote: Package: initramfs-tools Version: 0.59b Severity: grave Arch: sparc hmm it may be serious, but for now i go for important. anyway you omitted lots of information. Yes, I know ;-) It is impossible to build a ramfs image on a HyperSPARC workstation. I have done a mistake... I cannot use a ramfs image (I don't make ramfs images to boot my SS20). please post the error when trying to build the initramfs? I cannot post any error because HyperSPARC is not stable enough with 2.6 kernel. I only use kernels from debian repository or kernel in deb package I use for tests : http://www.wooyd.org/debian/kernels/linux-image-2.6.16-1-sparc32_2.6.16-6_sparc.deb I have try to install etch/sparc on a SS20 that runs with four HyperSPARC processors. When it boots, system hangs when it tries to load the ramfs image. please post the relevant messages. are you dropped into a console? No. Only a panic message with bad magic number when kernel tries to mount ramfs image. are your devices created? If I replace HyperSPARC by SuperSPARC-II, I can restart the workstation without any trouble. I have tested some configurations and some differents sparc32 CPU's. The only combinaison that fails is ramfs created by initramfs-tools with HyperSPARC. trave11er notified me of some sparc trouble, but from aboves bug report i have no idea where they could come from. For these tests, I have installed two SuperSPARC-II in place of ROSS modules. I use a patched kernel (ESP and SMP patches): hilbert:~# uname -a Linux hilbert 2.6.17-rc1-patch-smp #4 SMP Tue Apr 18 16:34:12 CEST 2006 sparc GNU/Linux hilbert:~# cat /proc/cpuinfo cpu : Texas Instruments, Inc. - SuperSparc-(II) fpu : SuperSparc on-chip FPU promlib : Version 3 Revision 2 prom: 2.25 type: sun4m ncpus probed: 2 ncpus active: 2 Cpu0Bogo: 75.16 Cpu2Bogo: 75.16 MMU type: TI Viking/MXCC contexts: 65536 nocache total : 5242880 nocache used: 1277440 State: CPU0: online CPU2: online hilbert:~# does fstype work? No, it does not... fstype is a sparcv9 executable ! hilbert:~# file /usr/lib/klibc/bin/fstype /usr/lib/klibc/bin/fstype: ELF 64-bit MSB executable, SPARC V9, version 1 (SYSV), statically linked (uses shared libs), stripped Even I kindly ask my SuperSPARC's, I think they cannot understand what they have to do... When I have installed initramfs-tools, I have installed the following packagesl: - busybox (1.01-4) - libklibc (1.2.4-1) - klibc-utils (1.2.4-1) - libvolume-id0 (0.089-1) - udev (0.089-1) - initramfs-tools (0.59b) please post output of: /usr/lib/klibc/bin/fstype /dev/sda1 # please use your real root cat /proc/cmdline hilbert:~# cat /proc/cmdline root=/dev/md1 ro md=1,/dev/sdb4,/dev/sda4 lsmod hilbert:~# lsmod Module Size Used by sg 33216 4 sr_mod 16680 0 cdrom 4 1 sr_mod sunlance 15024 0 openprom8560 0 openpromfs 17840 1 softdog 6644 0 hilbert:~# Don't forget kernel panics before it starts init. merci beacoup pour vos explications. sinon il m'est impossible de savoir ce qui ce passe. sincèrement Amicalement, JKB