1467764916 J * fstd_ ~fstd@x4e304a0a.dyn.telefonica.de 1467764916 Q * fstd Read error: Connection reset by peer 1467764933 N * fstd_ fstd 1467782770 Q * gamingrobot resistance.oftc.net beauty.oftc.net 1467782770 Q * jrklein_ resistance.oftc.net beauty.oftc.net 1467782770 Q * webhat resistance.oftc.net beauty.oftc.net 1467782797 J * webhat ~quassel@31.25.99.5 1467782826 J * gamingrobot sid10990@2604:8300:100:200b:6667:1:0:2aee 1467782932 J * jrklein ~cloud@proxy.dnihost.net 1467784847 J * Ghislain ~aqueos@adsl1.aqueos.com 1467786081 M * Bertl_oO Guy-: how exactly do you trigger the BUG in __release_vx_info()? 1467786963 M * Ghislain hi 1467786981 M * Ghislain i remember him saying doing /bin/sleep 1 onthe guest did it 1467787004 M * Bertl_oO yeah, but that is very unlikely 1467787011 M * Ghislain :) 1467787050 M * Ghislain well if the cpu does not feal like sleeping it can take this as an insult to his processing capabilities 1467787070 M * Ghislain feel 1467787089 M * Bertl_oO the problem is, the BUG() is reported in the release procedure, i.e. when a context (guest) terminates 1467787111 M * Bertl_oO I don't think that the guest will terminate on sleep, unless it is a very bad sleep :) 1467787222 M * Ghislain lol ! 1467788599 M * Ghislain on my side i was unable to do any errors but i only run sysbench and other in // 1467788633 M * Ghislain we will need Guy-: to tells us his workload 1467788653 M * Ghislain If you want me to test something just shoot... 1467789342 M * Guy- Bertl_oO: by running sleep(1) in a guest 1467789370 M * Guy- Bertl_oO: and no, the guest doesn't terminate 1467789401 M * Guy- what can I do to narrow it down further? maybe the vserver patch got mis-applied (there were some offsets, but I don't recall fuzz) 1467790181 Q * derjohn_mob Ping timeout: 480 seconds 1467790289 M * Ghislain i did not get any fuzz for 4.1.27 on my side 1467791498 J * _NaN_ ~croute@80.215.203.218 1467791735 J * derjohn_mob ~aj@fw.gkh-setu.de 1467799236 Q * _NaN_ Read error: Connection reset by peer 1467799750 Q * derjohn_mob Ping timeout: 480 seconds 1467800253 J * derjohn_mob ~aj@fw.gkh-setu.de 1467804888 M * Bertl_oO Guy-: so you ssh into your guest, and then execute /bin/sleep correct? 1467805055 M * Bertl_oO can you upload a minimal guest which causes this? i.e. only the necessary binaries and libraries to get it running 1467805496 M * Guy- Bertl_oO: no, I don't ssh in; there is a program that gets started on boot that periodically invokes sleep 1467805634 M * Guy- Bertl_oO: however, sleep(1) is only the most common process to trigger it; I have here also similar stacktraces from postmaster (psql), smtp (probably nullmailer), gzip 1467805688 M * Guy- now that I stopped the guest with the periodic sleep(1), the top offenders are: 1467805704 M * Guy- 7 gzip Tainted: P W O 4.1.27-vs2.3.8.4-starlifter #2 1467805704 M * Guy- 17 nmbd Tainted: P W O 4.1.27-vs2.3.8.4-starlifter #2 1467805704 M * Guy- 22 svn Tainted: P W O 4.1.27-vs2.3.8.4-starlifter #2 1467805704 M * Guy- 165 smbd Tainted: P W O 4.1.27-vs2.3.8.4-starlifter #2 1467805704 M * Guy- 246 smtp Tainted: P W O 4.1.27-vs2.3.8.4-starlifter #2 1467805707 M * Guy- 248 postmaster Tainted: P W O 4.1.27-vs2.3.8.4-starlifter #2 1467805803 M * Bertl_oO the kernl is tainted? 1467805814 M * Bertl_oO *kernel 1467805835 M * Bertl_oO can you ssh into a guest and invoke sleep to see if that triggers it? 1467805844 M * Guy- yes, the taint is due to zsh 1467805847 M * Guy- *zfs 1467805850 M * Guy- trying ssh 1467805862 M * Bertl_oO so zfs patch is involved as well? 1467805893 M * Bertl_oO what else did you add to the mix? 1467805987 M * Guy- zfs is not a patch, just an out of tree module 1467805990 M * Guy- there is nothing else 1467806027 M * Guy- and I have zfs on many boxes, but not this problem; also, this box did have zfs but not this problem with the previous kernel (which was also 4.1.something) 1467806049 M * Guy- I can't ssh in for some almost certainly unrelated reason, but vserver enter, then sleep triggers the BUG 1467806050 M * Bertl_oO so the kernel is only patched with a Linux-VServer patch? 1467806053 M * Guy- yes 1467806083 M * Bertl_oO okay, please be so kind and upload a (ideally minimal) guest 1467806111 M * Bertl_oO as well as your kernel .config 1467806115 M * Guy- OK, ssh worked, and sleep via ssh also triggers the BUG 1467806130 M * Guy- Bertl_oO: my .config is this: http://sprunge.us/HbcN 1467806140 M * Guy- I'll try to create a minimal guest 1467806291 M * Bertl_oO thanks! 1467807083 M * Guy- OK, I have a minimal guest 1467807221 M * Guy- it's a 20M tar.bz2 1467807229 M * Guy- how to best share it? 1467807496 M * Guy- Bertl_oO: https://files.cae-engineering.hu/sleep-bug.tar.bz2 1467807529 M * Guy- created using vserver sleep-bug build -m debootstrap -- -d jessie -s /usr/share/debootstrap/scripts/jessie -m http://http.debian.net/debian 1467808577 M * Bertl_oO 403 Forbidden 1467809073 M * Guy- oh, hang on 1467809126 M * Guy- Bertl_oO: can you try curl? the server insists on SNI, maybe wget doesn't send that 1467809164 M * Guy- someone from finland already downloaded it using weechat :) 1467809210 M * Bertl_oO okay, seems to work with curl 1467809370 M * Ghislain who use a browser anyway , command line is the only reality that counts 1467809373 M * Ghislain :) 1467809811 M * Bertl_oO indeed :) 1467810537 M * Bertl_oO Ghislain: can you test if the following patch works for you please? 1467810544 M * Bertl_oO http://vserver.13thfloor.at/Experimental/patch-4.1.27-vs2.3.8.5.2.diff 1467811873 J * Gremble ~Gremble@cpc87179-aztw31-2-0-cust6.18-1.cable.virginm.net 1467812468 M * Ghislain yep let me a little time this is customer phone day ! 1467812615 M * Bertl_oO isn't that usually a monday? :) 1467812751 M * Ghislain murphy's law can trigger this anytime you have something else to do 1467813705 M * Ghislain ok compilation is running, should take 20m then install reboot etc.. 1467813709 M * Ghislain keep you informed 1467813905 M * Guy- "TP-Link forgets to Renew and Loses its Domains Used to Configure Router Settings (tplinklogin.net)" /o\ 1467814748 M * daniel_hozac lol 1467814917 M * Ghislain well just throw them away and buy new ones, all IoT thing require the company that sells them to provide access, and 80% of them die in the 5 years so anyway.. even long standing company use that 1467814951 M * Ghislain i remember i could not configure my razer mouse because.. i had no internet access and it require to login razer portal to setup the mouse ! the freaking mouse 1467814984 M * Ghislain this all cloud thing is just enslaving people in every way they can 1467815683 M * Bertl_oO honestly, I wouldn't use such a device anyway 1467815813 Q * Gremble Quit: I Leave 1467816738 M * Ghislain well all are not telling you that they need internet access to work, the mouse for exemple i would never imagine that ( now they allow local settings after uproar of the comunity but see how they think..) 1467816758 M * Ghislain even cisco made switch that require cisco portal to work you could not access them directly 1467816784 M * Ghislain beware soon your fridge would not open if internet is down or if the gouvernement think you are too fat 1467820925 Q * derjohn_mob Ping timeout: 480 seconds 1467822631 J * derjohn_mob ~aj@88.128.80.62 1467823114 Q * sannes Ping timeout: 480 seconds 1467823675 J * sannes ~ace@2a02:fe0:c131:9070:a1c5:f88b:2534:b7eb 1467827064 Q * derjohn_mob Ping timeout: 480 seconds 1467834232 J * derjohn_mob ~aj@x590e37e1.dyn.telefonica.de 1467837853 J * derjohn_mobi ~aj@x590c58db.dyn.telefonica.de 1467838091 Q * derjohn_mob Ping timeout: 480 seconds 1467838365 J * BobR odie@IRC.13thfloor.at 1467838543 Q * BobR 1467847140 M * Ghislain kernel boot ok and seems to work, cannot test more now, basic mem limits works, go to test the tricky ones 1467847147 Q * Ghislain Quit: Leaving.