1578467610 M * Bertl_oO  off to bed now ... have a good one everyone!
1578467612 N * Bertl_oO Bertl_zZ
1578468556 J * Ghislain ~Ghislain@adsl1.aqueos.com
1578474409 J * hijacker ~nikolay@external.oldum.net
1578497654 N * Bertl_zZ Bertl
1578497657 M * Bertl  morning folks!
1578498727 J * x1450 ~1450@00027bbf.user.oftc.net
1578498792 M * x1450 Hi, I'm currently on Kernel 4.9.96-vs2.3.9.7 and having an issue with Monit - https://bitbucket.org/tildeslash/monit/issues/385/used-memory-size-detection - it appears that it uses SReclaimable which in VServer is passed from the host
1578498808 M * x1450 The patch that LXC implemented was to stub those values to 0 as per: https://www.redhat.com/archives/libvir-list/2016-January/msg00932.html
1578498831 M * x1450 I don't suppose this has already been fixed in vserver? :)
1578498986 M * x1450 if not, would someone be able to guide me to where i could stub this myself?
1578499361 M * Bertl hello x1450! sounds unusual ... let me read the links first :)
1578499509 M * Bertl ah, I see ... I presume you did enable the memory virtualization in your tests?
1578499546 M * Bertl (i.e. MemTotal (et al) are not the values you get from the host?
1578499806 J * fstd ~fstd@xdsl-89-1-52-208.nc.de
1578499981 M * x1450 sorry let me just double check
1578500014 M * x1450 I have the VIRT_MEM flag set
1578500047 M * x1450 so yeah the values in there are not the same as the host
1578500056 M * Bertl and a memory cgroup assigned to the guest (i.e. some kind of guest limits)
1578500075 M * x1450 no limits are set
1578500085 M * x1450 for memory
1578500114 M * Bertl hmm, so when you compare host and guest results, do they look different?
1578500233 M * Bertl btw, you probably want to update to 4.9.207+
1578500251 M * Bertl (not memory virtualization related though :)
1578500272 Q * fstd_ Ping timeout: 480 seconds
1578500317 M * x1450 https://pastebin.com/H2qnydrq
1578500361 M * x1450 Host (left) and Guest (right)
1578500391 M * x1450 I am planning on upgrading however if this isn't sorted in the latest kernel, I'd like to get it patched first as I can't do a deploy then a follow up deploy
1578500856 M * Bertl well, the relevant place is in fs/proc/meminfo.c
1578500873 M * Bertl (at least for the procfs entry)
1578500893 M * Bertl around line 113
1578500942 M * Bertl and you probably want to do a check for VIRT_MEM there and if set, replace the global_page_state() values with whatever you like (e.g. 0)
1578501566 M * x1450 thanks :)
1578501599 M * x1450 I'll give it a shot
1578501680 M * Bertl if you create a patch/diff, I'm happy to take a look :)
1578501695 M * x1450 If I can't manage it, I'll just work around it with a Monit patch although I'd prefer not to maintain a build for that :|
1578501718 M * x1450 appreciated
1578501742 M * Bertl something like:
1578501751 M * Bertl if (vx_flags(VXF_VIRT_MEM, 0)) {
1578501757 M * Bertl    virtualized stuff
1578501763 M * Bertl } else {
1578501768 M * Bertl    existing stuff
1578501770 M * Bertl }
1578501775 M * Bertl should do the trick
1578501891 M * Bertl note that VIRT_MEM is designed to work with memory limits, so in general it might be good to set some limits, even if they are not really limiting the guest
1578502567 M * x1450 ahh, I mainly just wanted the guest to have a unique view of it's memory space
1578502647 M * x1450 and and be able to set limits in the future if required :)
1578502687 M * x1450 just waiting for my kernel to build now and i'll hopefully be able to report back/provide a diff
1578502692 M * Bertl yeah, that's fine
1578505077 Q * hijacker 
1578506078 M * x1450 Any reason I wouldn't do it like so? https://pastebin.com/7wn34k4U
1578506087 M * x1450 it seems to work as expected
1578506128 M * x1450 produces: https://pastebin.com/GiPXRvb7
1578506374 M * Bertl you can do it this way, but it will be less performant
1578506400 M * Bertl i.e. the vx_flags(VXF_VIRT_MEM, 0) is the most expensive part here performance wise
1578506773 Q * obeardly 
1578507071 M * x1450 ahh ok thanks
1578507101 M * x1450 I'll refactor it into an if/else on Monday
1578507118 M * x1450 as I need to do it on a newer kernel and apply the rest of patches :)
1578507120 M * Bertl yeah, shouldn't be critical though
1578507346 M * x1450 is it something you'd consider rolling into core (I should submit a patch?) or no?
1578507739 M * Bertl it is not very invasive and if there are no unexpected side-effects I would consider including it
1578515880 Q * Aiken Ping timeout: 480 seconds
1578516260 J * Aiken ~Aiken@b951.h.jbmb.net
1578518135 M * Bertl  off for now ... bbl
1578518137 N * Bertl Bertl_oO
1578519063 Q * x1450 Ping timeout: 480 seconds
1578519146 J * DelTree ~deplagne@2a00:c70:1:213:246:56:18:2
1578519451 Q * Ghislain 
1578525217 J * Ghislain ~Ghislain@adsl1.aqueos.com
1578526751 Q * Ghislain