1515715369 J * Riviera Riviera@2a03:b0c0:1:d0::10:b001 1515718476 Q * guerby Remote host closed the connection 1515718487 J * guerby ~guerby@ip165.tetaneutral.net 1515720126 Q * guerby Remote host closed the connection 1515720199 J * guerby ~guerby@ip165.tetaneutral.net 1515721836 Q * guerby Read error: Connection reset by peer 1515721843 J * guerby ~guerby@ip165.tetaneutral.net 1515723606 Q * guerby Remote host closed the connection 1515723613 J * guerby ~guerby@ip165.tetaneutral.net 1515723756 Q * guerby Remote host closed the connection 1515723763 J * guerby ~guerby@ip165.tetaneutral.net 1515725586 Q * guerby Remote host closed the connection 1515725593 J * guerby ~guerby@ip165.tetaneutral.net 1515730146 J * guerby_ ~guerby@ip165.tetaneutral.net 1515730146 Q * guerby Read error: Connection reset by peer 1515732330 M * Bertl_oO off to bed now ... have a good one everyone! 1515732334 N * Bertl_oO Bertl_zZ 1515733335 J * fstd_ ~fstd@xdsl-85-197-45-69.netcologne.de 1515733782 Q * fstd Ping timeout: 480 seconds 1515733783 N * fstd_ fstd 1515738168 J * fstd_ ~fstd@xdsl-85-197-45-69.netcologne.de 1515738273 Q * fstd Remote host closed the connection 1515738274 N * fstd_ fstd 1515738604 J * obeardly ~obeardly@50-240-103-34-static.hfc.comcastbusiness.net 1515744988 Q * jrayhawk Quit: leaving 1515747183 J * jrayhawk ~jrayhawk@nursie.omgwallhack.org 1515758993 N * Bertl_zZ Bertl 1515759003 M * Bertl morning folks! 1515759082 M * gnarface morning Bertl! 1515764226 J * romster_ ~romster@158.140.215.184 1515764507 Q * romster Ping timeout: 480 seconds 1515766684 Q * Aiken Remote host closed the connection 1515769267 J * Gremble ~Gremble@cpc1-aztw34-2-0-cust397.18-1.cable.virginm.net 1515769842 M * Ghislain hello there, any need to help test anything today ? :) 1515769847 J * vanett_ ~vanett_@8L3AAAXC9.tor-irc.dnsbl.oftc.net 1515769943 P * vanett_ 1515769969 M * Gremble good to see the 4.4 patches out 1515769974 M * Gremble how are they looking? 1515770370 M * Le_Coyote Ghislain: while you're here, is it safe to enable COWBL in the latest patch Bertl? 1515770385 M * Le_Coyote I read that it has a bug, but will it break stuff if I don't use Cow/hashification? 1515771265 M * Ghislain well 4.4 seems to work on my test server i started to put it on several machine to see how this goes 1515771281 M * Ghislain i do not use Cow so i disabled it until further notices 1515771347 M * AlexanderS Bertl: I would like to see: https://gist.github.com/AlexanderS/6506e3d8b71c6bee8938b6faba39c175 and https://gist.github.com/AlexanderS/bd65ca5ffae0e11604f0272ddc53a03b and https://gist.github.com/AlexanderS/0665b4405d5a31a6406def2231ed46bc inside the "official" patches. Can I do anything to support it? 1515771557 M * Bertl you can start by explaining their purpose 1515771794 M * AlexanderS The first two patches are for enabling writable proc entries. The first one is a flag per proc entry and the second one (on top of the first one) adds a global flag for the guest. 1515771869 M * AlexanderS The last patch is simply for allowing mounting overlayfs inside a guest without the BINARY_MOUNT capability. 1515771990 M * Bertl well, with the first two I don't see a big issue 1515772023 M * Bertl the third one is problematic, because this can easily end up as an exploit 1515772136 M * AlexanderS Overlayfs only "merges" directories. So it is more like a sophisticated bind mount. 1515772405 M * Bertl yes, but being able to mount that inside a guest might cause all kind of issues 1515772423 M * Bertl starting from resource starvation to crashing the kernel 1515772644 M * AlexanderS The patch only removes the requirement of the BINARY_MOUNT capability. The SECURE_MOUNT capability is still required. 1515772995 M * Ghislain Bertl: not important but just because you are in full patching mode, the dir_emit is one thing but the dentry count is also not working ( increment to infinity) 1515773335 M * Bertl yeah, I remember 1515773676 M * Bertl I'm more worried about the cow link breaking at the moment, the dir_emit() can be considered resolved 1515774287 M * Ghislain yes, it was just because i stumbled on it today looking at the limits :) 1515774302 M * Ghislain like you know i dont use dentry limits either 1515774321 M * Ghislain so it just cosmetic 1515775051 M * Ghislain the dir_emit() can be considered resolved: i can add on my CV: patch accepted in the vserver kernel ;p 1515775101 M * Ghislain got to print a Tshirt 'kernel developper', and on the back 'bazingaaaaa' 1515775169 M * Bertl well, actually I threw the entire checks out, as you will see in the next patch :) 1515775199 M * Ghislain dam ! well to bad for this one then ;p 1515775524 Q * nikolay Quit: Leaving 1515775824 M * Le_Coyote Hm 1515775852 M * Le_Coyote Needs further testing but, for some reason, my first boot with 4.4.111 resulting in one of the vservers not starting at boot 1515775871 M * Le_Coyote Don't see how it could be connected though 1515775899 M * Bertl it could check something in /proc or depend on a kernel interface which is now disabled 1515775930 M * Le_Coyote Ok, second reboot, it's all back up. Weird. 1515776281 M * Le_Coyote Bertl: "free" has a weird behaviour inside the guest 1515776294 M * Le_Coyote It reports 0M used. I'm not sure it can be *THAT* efficient :) 1515776337 M * Le_Coyote mx1 ~ # free -m 1515776339 M * Le_Coyote total used free shared buff/cache available 1515776341 M * Le_Coyote Mem: 2048 0 2047 0 14 2047 1515776343 M * Le_Coyote Swap: 1024 0 1024 1515776345 M * Le_Coyote mx1 ~ # cat /proc/meminfo 1515776347 M * Le_Coyote MemTotal: 2097152 kB 1515776349 M * Le_Coyote MemFree: 2097116 kB 1515776351 M * Le_Coyote MemAvailable: 2238964 kB 1515776392 M * Le_Coyote top reports that mysql should at least be using 5% of that total 1515776408 M * Le_Coyote But MemAvailable being > MemTotal doesn't help, I guess 1515776415 M * Le_Coyote It's probably just cosmetic, but it's there 1515776457 M * Bertl cgroup active and configured properly? 1515776491 M * Le_Coyote I set a maximum mem and memswap 1515776496 M * Le_Coyote and VIRT_MEM for all the guests 1515776509 M * Le_Coyote It was working well in 3.18 a few minutes ago 1515776609 M * Bertl okay, kernel .config didn't change either? 1515777097 M * Le_Coyote The only VSERVER setting I changed was COWBL 1515777114 M * Le_Coyote Otherwise, I made oldconfig and didn't change anything "important" I think 1515777119 M * Ghislain i can also confirm the memory reported bu free -k is 312k here with apache mysql and virtualmin so there is perhaps an issue 1515777150 M * Le_Coyote 8k here 1515777172 M * Le_Coyote Thanks 1515777249 M * Le_Coyote More on this tomorrow, gotta run. Cheerio! 1515777265 M * Bertl tx, cya! 1515777316 M * Ghislain https://pastebin.com/raw/heWUSkHe, host run chrony and sshd only 1515777330 M * Ghislain well and rsyslog and cron 1515777393 M * Ghislain 1384952 with guest, 491836 after guest shutdown, 312 reported in guest 1515777408 M * Le_Coyote So you've got more free memory in the guest than in the host 1515777410 M * Le_Coyote Interesting 1515777414 M * Le_Coyote (yeah I'm almost going) 1515777430 M * Ghislain no its the used memory 1515777452 M * Le_Coyote "free" memory in the guest is higher 1515777467 M * Le_Coyote 16410316 > 15042440 1515777494 M * Le_Coyote There's already something a little fishy about that in 3.18 1515777500 M * Ghislain the guest use 893116k but report 312k used 1515777510 M * Le_Coyote htop reports I have 16ZETTAbytes occupied in the guest, out of 2G available 1515777513 M * Le_Coyote (yes ZETTA) 1515777559 M * Le_Coyote Really going now. See you tomorrow :) 1515777563 M * Bertl spendthrift 1515777613 M * Ghislain oh yes:  Mem[|                                                                                                16.0Z/15.7G] 1515777613 M * Ghislain same here ;p 1515777640 M * Ghislain curiously top is not having issue 1515777655 M * Ghislain iB Mem : 16410612 total, 16410296 free,      316 used 1515777655 M * Ghislain apart the low used mem 1515777712 M * Ghislain ==> memory.limit_in_bytes <== 1515777712 M * Ghislain 12884901888, so it should not show 16G but 12GB 1515777728 M * Ghislain well there is an issue tell me if you want me to test anything 1515777793 M * Bertl most likely this is a page vs byte count issue (or similar) 1515777816 M * Bertl you can test the raw values in /proc (not the interpreted ones) 1515777817 M * Ghislain sounds like it 1515777823 M * Bertl and compare them between host and guest 1515777839 M * Ghislain but there is alos the vmem that seems not to show 12gb 1515777841 M * Ghislain yes 1515779080 M * Ghislain sorry missed your sentence: https://pastebin.com/raw/fFwpR70V 1515779249 M * Ghislain https://pastebin.com/raw/50wZceaM diff between host and guest 1515779361 M * Ghislain memfree and memavailable are > in guest than in host 1515781112 M * Bertl ah, finally, cow link breaking seems to be fixed 1515781124 M * Bertl will upload an updated patch shortly for testing 1515781390 M * Bertl http://vserver.13thfloor.at/Experimental/patch-4.4.111-vs2.3.9.3-testme.diff 1515784711 Q * arekm Remote host closed the connection 1515784939 M * Bertl off for now ... bbl 1515784944 N * Bertl Bertl_oO 1515785699 J * Aiken ~Aiken@2001:44b8:2168:1000:b26e:bfff:fe2a:b951 1515785879 Q * Gremble Quit: Leaving 1515787704 Q * ntrs Read error: Connection reset by peer 1515799203 Q * Guy- Ping timeout: 480 seconds 1515801440 J * Guy- ~korn@elan.rulez.org