1518135741 M * Bertl_oO off to bed now ... have a good one everyone! 1518135742 N * Bertl_oO Bertl_zZ 1518160677 N * Bertl_zZ Bertl 1518160682 M * Bertl morning folks! 1518160903 Q * webhat Quit: http://quassel-irc.org - Chat comfortably. Anywhere. 1518160917 J * webhat ~quassel@31.25.99.5 1518166228 M * gnarface morning Bertl! 1518167065 J * nikolay ~nikolay@149.235.255.3 1518172913 M * Bertl off for now ... bbl 1518172918 N * Bertl Bertl_oO 1518179260 Q * Aiken Remote host closed the connection 1518191954 M * Ghislain1 humm i have vcontext Zombies, any idea how to debug this one ? 1518191954 M * Ghislain1 root     18258  0.1  0.0      0     0 pts/0    Z    16:54   0:00 [vcontext] 1518191985 M * Ghislain1 cmdline, exec cwd are all empty 1518192271 M * Bertl_oO with or without the exit-fix? 1518192602 M * Ghislain1 no this is on old 4.1 kernel not the latests ones 1518192787 M * daniel_hozac when do you get them? who is the parent? 1518192895 M * Bertl_oO 4.1.x is also affected by the exit problem 1518193069 M * Ghislain1 ps   -o ppid= 18258 1518193070 M * Ghislain1 18257 => does not exist 1518193094 M * Ghislain1 oh wait 1518193102 M * Ghislain1 it exsit on the host this is a vserver enter 1518193141 M * Bertl_oO mystery solved :) 1518193153 M * Ghislain1 hum should appear when stopping a guest while i was in a vserver enter on another window 1518193164 M * Ghislain1 cannot kill it 1518193183 M * Bertl_oO it's already dead, it just didn't get reaped 1518193211 M * Ghislain1 yes, visibly kill have not worked but after a while it disapeared so we are back on feet 1518193621 M * torrunes wonder if there is some leakage in 4.9 patches... Is it correct that within a vs, netstat can show port statuses from other vs's and the host: 1518193630 M * torrunes sg1 ~ # vserver pfadata enter 1518193642 M * torrunes pfadata / # netstat -atnp 1518193643 M * torrunes Active Internet connections (servers and established) 1518193645 M * torrunes Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name 1518193646 M * torrunes tcp 0 0 127.0.3.1:3306 0.0.0.0:* LISTEN - 1518193647 M * torrunes tcp 0 0 10.0.0.201:3306 0.0.0.0:* LISTEN - 1518193649 M * torrunes tcp 0 0 10.0.0.10:2222 0.0.0.0:* LISTEN - 1518193650 M * torrunes tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN - 1518193652 M * torrunes tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN - 1518193653 M * torrunes tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN - 1518193655 M * torrunes tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN - 1518193656 M * torrunes tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN 12962/sshd 1518193658 M * torrunes tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN - 1518193659 M * torrunes tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN - 1518193661 M * torrunes tcp 0 0 0.0.0.0:443 0.0.0.0:* LISTEN - 1518193662 M * Ghislain1 whoha 1518193662 M * torrunes tcp 0 0 0.0.0.0:443 0.0.0.0:* LISTEN - 1518193667 M * torrunes the ip of this vs is 10.0.0.169 1518193672 M * Ghislain1 dont do that plz 1518193682 M * Bertl_oO yeah, please use pastebin or similar 1518193684 M * Ghislain1 use pastebin 1518193685 M * torrunes sry 1518193687 M * Ghislain1 :) 1518193718 M * torrunes quite new to this - feel free to kick my butt... 1518193768 M * Ghislain1 ohoh i confirm that 1518193796 M * Ghislain1 netstat report them, i do not see the program (-p) but i definitively see other guest open ports 1518193818 M * torrunes i dont have the prev 4.1 patches at hand right now, but I don't think that was the case in the older patches...(?) 1518193845 M * Bertl_oO are the ports bound to 0.0.0.0 or are they bound to other guest IPs? 1518193847 M * Ghislain1 no, old 4.1 do not show them 1518193864 M * Ghislain1 on 0.0.0.0 1518193876 M * Ghislain1 tcp        0      0 0.0.0.0:25              0.0.0.0:*               LISTEN      2057/master 1518193876 M * Ghislain1 tcp        0      0 0.0.0.0:25              0.0.0.0:*               LISTEN      - 1518193887 M * torrunes at least some of the 0.0.0.0 are from other guests 1518193905 M * Ghislain1 oh drn both ! 1518193906 M * Bertl_oO can you test with a specifically IP bound port from a different guest? 1518193913 M * Ghislain1 i also see the others ip in them 1518193927 M * Bertl_oO okay, then the check is missing somewhere 1518193977 M * Ghislain1 it show also the 127.156.254.1 of the other guest 1518194025 M * Ghislain1 good catch torrunes, thats what i get for testing with one guest only 1518194044 M * Bertl_oO so the next step is to figure out where netstat gets the data from 1518194152 M * torrunes Bertl, e.g. the line tcp 0 0 10.0.0.201:3306 is the mysql server of another guest. To test it, I stopped it, and it was gone from the other guest. 1518194276 M * torrunes if you want me to test something, just ping here 1518194293 M * Bertl_oO best use strace -fF to see where netstat gets the data from 1518194306 M * Bertl_oO probably /proc or /sys 1518194334 M * torrunes on it 1518194392 M * daniel_hozac IIRC /proc/net/tcp 1518194414 M * torrunes could it be that proc security is not functioning? 1518194416 M * daniel_hozac does ss -nl also show other guests? 1518194474 M * Bertl_oO and does it happen with 4.4.x as well or 'just' 4.9.x? 1518194490 M * Ghislain1 it shows the internal 127.X.X.X ip ss -nl 1518194500 M * torrunes ls /proc within guest shows a bit more than I usually expect. 1518194505 M * torrunes have not tried 4.4 1518194509 M * Ghislain1 but not the other ip 1518194545 M * torrunes need to find an older 4.1 to check. BRB 1518194549 M * Ghislain1 i only have a recent 4.9, i quite abandonned the 4.4 line since the 4.9 is here i must admit 1518194569 M * Bertl_oO can you test with 4.4.x as well, just to make sure? 1518194592 M * Ghislain1 i must recompile a 4.4 so it will take some thime 1518194617 M * Bertl_oO no problem, tx 1518194681 Q * nikolay Remote host closed the connection 1518194797 M * torrunes ls /proc seems ok. False alarm, sry 1518194931 M * torrunes strace -fF netstat -atpn within guest: https://pastebin.com/KiNdhiMV 1518195011 M * Bertl_oO so it seems to be /proc/net/tcp as daniel_hozac suspected 1518195127 M * Ghislain1 yes i cannot really read it as human but the size is 2x less if i stop a guest so... 1518196755 M * Bertl_oO let's see if that fixes it: 1518196757 M * Bertl_oO http://vserver.13thfloor.at/Experimental/delta-ipv4-fix01.diff 1518197191 M * torrunes will test during weekend ! :) 1518197915 M * Ghislain1 4.4.115 dos not seems to leak the sockets 1518198008 M * Ghislain1 4.4.115-vs2.3.9.6aq does not leek the sockets 1518198125 M * Ghislain1 4.4 had vxdprintk(VXD_CBIT(net, 6), "sk: %p [#%d] (from %d)",sk, sk->sk_nid, nx_current_nid()); 1518198125 M * Ghislain1 also 1518198364 M * Bertl_oO yeah, I'll add that back as well 1518199559 M * Bertl_oO here it is ... http://vserver.13thfloor.at/Experimental/delta-ipv4-feat01.diff 1518201274 J * wonder1fu ~ipfvmn@45.76.243.90 1518201318 M * wonder1fu (_) | | 1518201326 M * wonder1fu _ _ __ ___ ___ _ _ _ __ ___ _ __ _ __ ___| |_ ___ ___ _ __ __ _ 1518201350 M * wonder1fu | | '__/ __| / __| | | | '_ \ / _ \ '__| '_ \ / _ \ __/ __| / _ \| '__/ _` | 1518201350 M * wonder1fu | | | | (__ _\__ \ |_| | |_) | __/ | | | | | __/ |_\__ \| (_) | | | (_| | 1518201350 M * wonder1fu |_|_| \___(_)___/\__,_| .__/ \___|_| |_| |_|\___|\__|___(_)___/|_| \__, | 1518201350 M * wonder1fu | | __/ | 1518201454 M * wonder1fu |_| |___/ 1518201454 M * wonder1fu webhat guerby_ PowerKe torrunes romster Ghislain1 daniel_hozac dustinm` ggherdov__ pulsar ntrs jrklein_ LongyanG zerick BWare Pinky Bertl_oO yang_ padde funnel DoberMann sladen yang transacid _are_ gratuxri Carpoon gnarface Guy- arekm AlexanderS Rockj bzed tokkee karasz kolorafa_ geb AndrewLe1 jrayhawk Jb_boin Le_Coyote DLange BlackPanx mnemoc kshannon clopez l0kit mcp CcxWrk DelTree Hunger _Shiva_ FloodServ 1518201454 Q * wonder1fu 1518201572 M * Bertl_oO wow, he knows how to 'use' figlet :) 1518203077 J * obeardly ~obeardly@12.153.3.34 1518204591 J * Aiken ~Aiken@2001:44b8:2168:1000:b26e:bfff:fe2a:b951