1516322963 Q * obeardly Quit: Leaving 1516322983 J * obeardly ~obeardly@12.153.3.33 1516324055 Q * obeardly Ping timeout: 480 seconds 1516324101 J * obeardly ~obeardly@12.153.3.34 1516332967 M * Bertl off to bed now ... have a good one everyone! 1516332969 N * Bertl Bertl_zZ 1516342987 Q * Aiken Remote host closed the connection 1516343040 J * Aiken ~Aiken@2001:44b8:2168:1000:b26e:bfff:fe2a:b951 1516348251 M * yang_ Bertl_zZ: its a problem with grub now, which doesnt install 1516348324 J * ksn ~ksn@00019501.user.oftc.net 1516351694 J * fstd_ ~fstd@xdsl-87-78-139-40.netcologne.de 1516351858 M * Guy- yes, grub-install says it can't recognize the filesystem, and we don't know why; blkid does recognize it 1516351876 M * Guy- (it's ext4, so nothing extreme) 1516352143 Q * fstd Ping timeout: 480 seconds 1516352143 N * fstd_ fstd 1516354135 J * nikolay ~nikolay@149.235.255.3 1516354744 N * Bertl_zZ Bertl 1516354761 M * Bertl morning folks! 1516354778 M * Bertl Guy-: is it grub or grub2? 1516354796 M * Guy- grub2 1516355372 M * Bertl what grub modules are available? 1516355435 M * Bertl and is there a separate /boot partition or just one big filesystem (the one you are trying to repair)? 1516355458 Q * ksn Read error: Connection reset by peer 1516355486 J * fstd_ ~fstd@xdsl-87-78-139-40.netcologne.de 1516355643 Q * fstd Remote host closed the connection 1516355643 N * fstd_ fstd 1516356659 J * ksn ~ksn@00019501.user.oftc.net 1516360023 Q * ksn Ping timeout: 480 seconds 1516360611 M * Guy- Bertl: it's one big fs unfortunately, and all grub modules should be available (we tried reinstalling the grub packages to make sure) 1516360638 M * Guy- e2fsck is able to fix the fs to the point where a 2nd invocation of e2fsk -f doesn't find anything wrong with it 1516360667 M * Guy- however, if we mount it (or some time after that), one of the entries under lost+found becomes corrupted 1516360680 M * Guy- e2fsck is able to fix it again, and then it becomes corrupted in the same way again 1516360713 M * Guy- I was hoping to be able to boot the initramfs of the real, live system, because it might have a different version of e2fsck than the rescue environment (not to mention a different kernel) 1516360720 M * Bertl sounds like a disk problem to me (or a very weird fsck/ext4 incompatibility) 1516360725 M * Guy- but the machine fails to boot, dropping to a grub rescue shell 1516360734 M * Guy- there are no block i/o errors in dmesg, though 1516360745 M * Bertl well, you can still boot from a grub rescue shell 1516360750 M * Guy- it could still be a disk problem -- I once had a disk where all writes silently failed 1516360765 M * Bertl what does 'ls' show on the grub shell? 1516360785 M * Guy- I don't know, we concentrated on fixing grub 1516360800 M * Guy- didn't experiment with the grub rescue shell much (other than trying 'normal', which failed) 1516360829 M * Bertl the e2fsck is from the rescue disk, yes? 1516360834 M * Guy- my next two ideas would be trying a different bootloader (say, syslinux) and biting off a chunk of swap for a separate /boot partition, see if that helps 1516360847 M * Bertl (which is probably some cd or so?) 1516360847 M * Guy- yes, the e2fsck is from the rescue environment 1516360864 M * Guy- I think it's probably netbooted, from an image provided by the hosting provider 1516360885 M * Guy- it has a 4.12.0 kernel and not the latest e2fsprogs 1516360903 M * Bertl strange combination 1516360923 M * Bertl anyway, the data was rescued with the rsync, no? 1516360938 M * Guy- yes, there is a backup (except the one corrupted directory from lost+found, I guess) 1516360952 M * Guy- which is the root of a vserver guest 1516360964 M * Bertl what's the point in hacking at the current image? 1516360993 M * Guy- two points: 1. if it were possible to fix, the day-long restore would not be necessary; 2. it might be possible to rescue that vserver guest 1516361023 M * Bertl for 1. you probably have spent more time on the 'fix' than a restore would have taken 1516361032 M * Guy- (3. it always helps to get to understand an issue, because it might help the next time you see something similar) 1516361046 M * Guy- yes, but that wasn't apparent in the beginning 1516361048 M * Bertl for 2. if data recovery is essential, you should have made a low level copy of the disk data 1516361062 M * Guy- I have no idea how essential it is, I'm just helping out 1516361063 M * Bertl (and analyze it in a clean environment) 1516361087 M * Bertl yeah, doesn't so much go in your direction, I'm just saying 1516361088 M * AlexanderS It might be easier to backup the whole disc to a local device (using f.e. dd and nc) and try to fix/analyse it locally first. 1516361091 M * Guy- my impression was it can't be too essential, because the box has been down for a week now and people are not on fire 1516361130 M * Bertl so probably not more important than lkml :) 1516361134 M * Guy- AlexanderS: there is no local drive; I think this is a rented server at a remote location -- you'd have to transfer several hunderd gigabytes to get an image you can analyze (doable, but slow) 1516361147 M * Bertl local drive would be 'at home' 1516361156 M * Guy- ah, local in that sense 1516361157 M * Guy- yes 1516365768 M * Ghislain1 wont to sound silly but tried spinrite on the disk ? 1516365912 J * druschka_domaintechnik ~druschka@85.118.185.154 1516367681 M * Bertl off for now ... bbl 1516367686 N * Bertl Bertl_oO 1516371051 Q * druschka_domaintechnik Quit: druschka_domaintechnik 1516373550 M * Ghislain1 is anyone ever succeeded to have memory.use_hierarchy 1516373550 M * Ghislain1 to 1 ? 1516373569 M * Ghislain1 lsvs 1516373575 M * Ghislain1 oups sorry 1516376419 Q * nikolay Quit: Leaving 1516379040 M * Ghislain1 Bertl_oO: how would you find in the kernel code where the cgroup system create /memory.stat ? 1516379873 Q * arekm Ping timeout: 480 seconds 1516380401 J * arekm ~arekm@phobos.pld-linux.org 1516383050 J * druschka_domaintechnik ~druschka@85.118.185.154 1516383273 J * Gremble ~Gremble@cpc1-aztw34-2-0-cust397.18-1.cable.virginm.net 1516384835 Q * Gremble Quit: Leaving 1516385501 M * Ghislain1 i think i have a clue about the virtmem, testing this after the kernel compile 1516386542 Q * druschka_domaintechnik Quit: druschka_domaintechnik 1516387172 Q * padde Remote host closed the connection 1516387449 J * padde ~padde@patrick-nagel.net 1516388546 M * yang_ Bertl_oO: Guy- explained it for you. The backup is somehow not top priority restore, as there are no active users waiting, except for me and my own data, which is important 1516388576 M * yang_ In case I don't manage to do a restore, I will most likely abadon the server there 1516388582 M * yang_ It's a dedicated server 1516394344 Q * gnarface Read error: Connection reset by peer 1516396267 J * gnarface ~gnarface@108-227-52-42.lightspeed.irvnca.sbcglobal.net 1516397678 M * Ghislain1 GUEST:root@stretch01:[~]: free -k 1516397678 M * Ghislain1              total        used        free      shared  buff/cache   available 1516397678 M * Ghislain1 Mem:       12582912     1760112    10803532       31468       19268    10719064 1516397687 M * Ghislain1 i seems to work ! 1516397729 M * Ghislain1 bertl_o0: my fix for the virtmen issue: vserver/limit.c 1516397735 M * Ghislain1        if (res_limit != PAGE_COUNTER_MAX) 1516397735 M * Ghislain1                val->totalram = res_limit * ( PAGE_SIZE / val->mem_unit) ; 1516397735 M * Ghislain1        val->freeram = val->totalram - ( res_usage * PAGE_SIZE / val->mem_unit ); 1516397762 M * Ghislain1 the sysinfo unit vary, you have to convert because this is not allways PAGESIZE 1516397774 M * Ghislain1 not on 64bit they directly put bytes 1516397784 M * Ghislain1 val-mem_unit =1 1516401277 M * Aiken Ghislain1, which kernel version are you playing with? Been toying with the idea of 4.9.76 on my vmserver 1516405709 J * itsme^ ~heybaby@7YZAAA68M.tor-irc.dnsbl.oftc.net 1516405997 P * itsme^