=== night is now known as night|pub === night|pub is now known as night [07:32] mornings [07:41] The last 3 days I've been getting "hash-sum mismatched" on xenial-updates for 3 computers that are configured to use either of the two mirrors: archive.ubuntu.com and ubuntu.mirror.neology.co.za . Curious as to whether anyone else running 16.04.x has also been experiencing those issues. [07:42] I've just done: apt clean ; sudo rm -r /var/lib/apt/lists/* -vf ; apt update and am still getting that error for Err:9 http://archive.ubuntu.com/ubuntu xenial-updates/main amd64 Packages [07:42] Hash Sum mismatch [07:52] hmm. and now after repeating the same set of "clean, rm, clean, update" the error has disappeared. Odd. I guess some file or apt-related process must have still been open [07:58] andrewlsd: FWIW I've been using mirror.wiru.co.za hassle-free for a few weeks instead of the neology mirror [07:58] also 1G uplink and seems to update more frequently [07:59] I've seen neology's mirror lag 2-3 days behind a few times recently [08:21] good morning all have a great day [08:27] So I have a few cloudatcost boex (yes I know they suck etc.) that I ran some non crucial stuff on, they mostly work okay but there is this problem where every now and then the filesystem becomes readonly and then the stuff stops working. [08:27] I figure the easiest fix is to just reboot the machines, what would be the best way to detect that a filesystem has gone read only? [08:28] cron job, or is there some more automatic way to have it reboot in such circumstances? [08:38] hi mani [08:38] if the reboot command work in terminal without asking for sudo password then you can set-up a cron job [08:56] MaNI: tune2fs -l might show if filesystem is currently read-only. [09:16] thanks thatgraemeguy for the mirror.wiru tip. [09:17] Anyone here know of any cape town-based individuals or (small) companies that I could refer a small organization to for Linux support. Basically a firewall+proxy setup? [09:18] where "know of" means you'd like to suggest them as worth talking to. rather than have just heard of someone. Thanks everyone in advance. [09:33] MaNI: look at monit, it's designed for that sort of thing [09:37] thanks, I'll see if I can find some text that describes it's capabilities [09:38] am I the only one that misses the days when webpages actually had information, instead of being a scroll fest with one line of marketing per overly graphic designed page [09:39] andrewlsd: way back when I used to use Clarotech Consulting. I haven't needed them in ages, but I see from their site the management team is still the same, so I reckon they're still just as good as ever [09:40] thanks thatgraemeguy, yeah I know some of the Clarotech people. Good recommendation. [09:43] monit does look good for the task, thanks [09:45] the real fix of course is to see why it does that. you probably have 'errors=remount-ro' in the mount options, so there is an underlying error causing it to remount the filesystem/s read-only [09:46] 'dmesg' might have clues, otherwise perhaps the host's support people can assist [09:46] yeah, I think it's out of my control [09:46] AFAICT the harddrive performance becomes so poor at times that the system does an emergency remount [09:48] better to let it do that and reboot than to risk filesystem corruption [09:52] or at least in this case, if it were for anything important I'd burn the servers and then get real ones :) [09:56] fair enough [10:01] MaNI: you might want to start remote logging, since if root filesystem goes R/O, on a cloud system, then you probably won't get logs written to disk. [10:01] or just get a second disk / partition just for logging to. [10:43] yeah maybe [10:55] MaNI: Can you do a `mount -o remount,rw /`? [10:55] Doesn't work on btrfs, not sure if that applies to all filesystems though [11:18] MaNI: probably btrfs is a _bad_ idea on unreliable storage. (but you probably realized that by now) [11:19] use case for an alpine ^ run in ram, and periodically sync (backup) configs and data to persistent disk. [11:20] yeah, it's ext4 [11:20] phew. ask theblazehen about fun with `btrfs`. [11:20] where s/fun/data\ loss/ [11:21] IIRC [11:21] I got over playing with filesystems years ago, hehe [11:23] MaNI: you could set the filesystem error behaviour to panic. That might give you a reboot when there are fs issues. instead of needing to detect RO and reboot. [11:23] hrm, yeah that would be handy if it works [11:29] okay, set now to wait a week or so and see what happens :) [11:31] you could test it by creating loopback device with that setting https://serverfault.com/questions/498900/intentionally-cause-an-i-o-error-in-linux/498931 [11:31] and manually generate a fs error [11:31] Yeah :( [11:31] At least, now I know snapshots != backups [11:34] I had some 'fun' experimenting with murderfs back in the day, and then I ran into a 'rare' jfs kernel bug [11:34] heh [11:35] after that some on and off issues with xfs, and I decided it's best to just stick with the most used/tested filesystems and not try anything different [11:35] I'm using btrfs with daily backups to ext4 [11:37] whats the SSD fs from samsung called? [11:37] F2FS. [11:38] Flash Friendly Filesystem [11:38] https://en.wikipedia.org/wiki/F2FS [11:38] oh yeah I also had some fun a few years ago experimenting with the various 'SSD cache' systems in the kernel, bcache etc. [11:39] my end conclusion was that none of them should be let anywhere near any data that matters :( [11:39] at least not with the SSD drive I was testing with, maybe people have better luck with other drives [11:39] * andrewlsd thinks MaNI and theblazehen reached a similar conclusion regarding RAM caches [11:43] Yeah, RAM is no. bcache works well though [11:48] Just don't use bcache on a file in /dev/shm [11:49] I was getting regular corruption in my tests, but this was a few years ago, and it may have been some specific issue with that particular SSD and/or xfs or something [11:50] the SSD itself has worked 100% reliably since then as a normal SSD, so it wasn't faulty or anything, but maybe still just some funny with it and bcache [11:52] Did you have discard enabled? [11:56] too long ago to answer honestly :/ but I vaguely recall I may have tried both with it on or off [11:58] I do also use zram, so it's possible theres some weird interaction there [13:20] Hi All, hows everyone doing? [13:35] anybody here know if Ubuntu will work on a Gigabyte Laptop? [13:37] thats going to depend very much on the model I would imagine [13:38] but generally would you say just the default install should work or do you suspects it might not be that easy? [13:44] probably, but laptops are fiddly when it comes to graphics drivers, bluetooth drivers, wireless drivers etc. so it's not uncommon for certain laptops to have problems [14:31] Hey guys! [14:32] Does anyone use WordPress' CLI tool to perform scheduled backups? [14:32] something like `wp export` or `wp db export` [15:45] propagandhi: I would just backup WordPress directory and export DB === MaNL is now known as MaNI [18:54] Evening :) [18:55] join #ubuntu-za [18:55] sorry [18:55] lol [18:56] evening calhax_za [18:57] hows you doing nsnzero? [18:57] where are kilos and inetpro? sleeping? [18:58] ok and yourself calhax_za [18:58] i havent seen either of them in a while [18:59] im good thanks nsnzero, I havent seen you here before, althou I haven't been around for a while LOL [19:01] well in that case nice to meet you - i joined fairly recently [19:02] awesome sauce nsnzero [19:52] evening all [21:39] ohi