libera/#devuan/ Thursday, 2023-08-17

systemdleteWill 20GB of disk space be adequate for a daedalus install?  This would exclude large applications, which I can install on other drives.00:54
bb|hcbshould be more than enough, imho :)00:55
systemdletebb|hcb, how much space does yours take?00:55
systemdleteI'm just trying to get a rough idea.00:56
systemdleteI usually figure about 10G, but I might want to add utilities and the like later on.00:56
brocashelmyou could add partitions for things like /tmp and /var so / doesn't get completely filled up00:57
brocashelm20 gb is fair00:57
bb|hcbThe one I am sitting on uses 118G, but that is 10+ years cruft (and there is no one to clean it)00:57
brocashelmyeah, mine is from my beowulf install circa 202000:58
systemdleteRight. that's what I have done in the past.   For chimaera I allotted 16GB00:58
brocashelmalthough i did a lot of cleanup00:58
brocashelmi recommend using ncdu to check which directories/files use up the most space00:58
systemdletencdu will not help me before I install it, though.  :)00:59
bb|hcb...and that was Debian in its past reincarnation, most files in one partition00:59
bb|hcbBTW. I have a laptop where the usage is 14G (excluding home, docs and downloads). If you keep big stuff on separate partitions, 20G is plenty01:00
brocashelmi meant after the install, so you can review storage01:00
brocashelmlike i said, 20 gb is far. 30 gb is ideal IMO01:01
brocashelm*fair01:01
systemdleteIn the past, for beowulf and chimaera, I figured 6GB for /usr, 2GB for /var, 2GB for /tmp, 4GB for /home, and the remainder for /01:01
systemdleteAnd I use lvm so I can extend those by adding more PVs01:01
brocashelmthinking it might also be useful to create a partition for /usr/local01:01
brocashelmfor programs you compile/install locally, so that if they run into a problem, they can't take your whole /usr down01:02
systemdletebrocashelm, thank you.  I figured 20GB would be enough, but for chimaera, someone said 30GB, which I did in one instance.  But I found that 20GB seemed enough for chimaera01:02
systemdletebrocashelm, I agree. totally!01:02
brocashelmsoftware has been getting increasingly heavier since chimaera (gtk3), so more resources is a plus01:03
systemdleteI don't do much with /usr/local, but for large applications, I allocate a separate fs and load each one in its own fs.01:03
systemdleteright, that's what I am concerned with.   I could not find doc on this at devuan or debian01:03
systemdlete(It's time for open source to go on a diet.)01:04
systemdleteSo I am off to attempt an install of daedalus...01:05
brocashelmjust going off my openbsd install, it comes with several partitions, and /usr/local is one of them01:05
brocashelmgood luck01:05
systemdletethanks01:05
brocashelmdebfoster FTW01:05
systemdletewho?01:06
systemdletejust searched.  There are 1500 people with that name.01:06
systemdleteone just had a baby by caesarian01:07
bb|hcbIMHO, with most of the stuff on the local drive being scrappable, there isn't much sense to do separate partitions - when all the code and documents are in some git, starting on a fresh install shouldn't be that hard?01:09
systemdletebrocashelm, did you mean deshaun foster for the win?01:11
systemdlete(sorry)01:11
systemdletebb|hcb, sounds like you are talking about a developers environment.  That might not work for others though.01:12
bb|hcbAh, maybe, I am talking for what I do... I even prefer not to customize the DE, because on each new machine, I will have to repeat that, and it is a hassle...01:15
brocashelmsystemdlete: if you read the log before signing back on: debfoster is a tool for checking which debian packages are the reason for having a lot of cruft installed (similar to deborphan, but debfoster goes through every main package on your system). i use it to cut down on bloat, only using the core stuff01:30
onefangetckeeper automatically keeps track of changes to /etc using git.03:24
rwp+1 for etckeeper!  etckeeper rocks.  FTW!05:05
rwpUsing etckeeper I almost never "remove" anymore and always "purge" because I know the /etc conffiles are safely stored and available in the git repository if I want them.05:05
adaNewb1234is anyone else having issues updating mariadb-server?08:02
gnarfaceadaNewb1234: not that i've seen anyone mention here recently, but it's likely that any issues would be common to Debian or even past versions of mysql-server (i remember a common known gotcha about corrupted tables; it should still be easy to find the documented fix)08:08
adaNewb1234The only diagnostics I have immediately available is dpkg saying: "Setting up mariadb-server (1:10.11.3-1) ...08:13
adaNewb1234Stopping MariaDB database server: mariadbd.08:13
adaNewb1234Starting MariaDB database server: mariadbd . . . . . . . . . . . . . . . . . . .08:13
adaNewb1234 . . . . . . . . . . . failed!08:13
adaNewb1234invoke-rc.d: initscript mariadb, action "start" failed.08:13
adaNewb1234dpkg: error processing package mariadb-server (--configure):08:13
adaNewb1234 installed mariadb-server package post-installation script subprocess returned e08:13
adaNewb1234rror exit status 108:13
adaNewb1234"08:13
gnarfacecheck the /var/log/daemon.log08:16
gnarfaceif a daemon failed to start it should say why08:16
adaNewb1234there's nothing there, interestingly enough.08:21
gnarfacehmm, how about in /var/log/mariadb?08:21
gnarfacemysql did have its own logs too, as i assume mariadb does, but you may have to enable them in the config first to get anything useful08:22
gnarfaceif it didn't get as far as running the daemon binary though it could be a problem in the init script or even the package post-install script itself. what release version of devuan are you running?08:23
gnarfaceif it's ceres, i'd just wait a day and try again08:24
gnarfaceif it's a earlier release i'd check the debian bug tracker to see if there's a bug in their corresponding version that's already been reported08:24
gnarfacealternately, if you just try to call the init script with the "start" parameter directly it might be obvious what's wrong08:25
adaNewb1234$ sudo service mariadb start08:29
adaNewb1234Starting MariaDB database server: mariadbd.08:29
adaNewb1234$ sudo service mariadb status08:29
adaNewb1234Copyright (c) 2000, 2018, Oracle, MariaDB Corporation Ab and others.08:29
adaNewb1234Server version10.11.3-MariaDB-108:29
adaNewb1234Protocol version1008:29
adaNewb1234ConnectionLocalhost via UNIX socket08:29
adaNewb1234UNIX socket/var/run/mysqld/mysqld.sock08:29
adaNewb1234Uptime:5 sec08:29
adaNewb1234Threads: 1  Questions: 87  Slow queries: 0  Opens: 46  Open tables: 39  Queries per second avg: 17.400.08:29
adaNewb1234interestingly enough08:29
adaNewb1234So, I guess the problem would be in the post-install script?08:30
adaNewb1234It works, but dpkg doesn't believe it.08:31
gnarfacewell it worked when you called it08:37
gnarfacemaybe there was a problem creating the socket file the first time or something?08:37
gnarfacei'm not sure but yea, check the post-install script to see if it's doing anything weird08:37
gnarfaceseems like it should be a known issue if it wasn't self-inflicted somehow08:38
adaNewb1234ok, so invoke-rc.d mariadb restart fails, but service mariadb restart doesn't.08:51
adaNewb1234don't they both call the same script?08:53
gnarfacehuh08:53
gnarfacenot sure honestly, i always just call /etc/init.d/[whatever] directly08:53
gnarfaceif "service mariadb..." was the one that failed i'd suspect a missing package (the one with the service script itself)08:54
gnarfacepossible invoke-rc.d also requires some package?08:54
gnarfacewell, yes i mean, ostensibly they're both calling something in /etc/init.d/mariadb via a symlink pointing to it from /etc/rc*.d/ but i couldn't guess through how many intermediaries08:57
gnarface*something like /etc/init.d/mariadb i mean08:57
gnarface(and i've seen even that part done wrong with something before too, where they are assuming the redhat default runlevel and the debian one had no "start" symlink there or something like that)08:58
gnarfaceadaNewb1234: if you figure it out, let us know09:14
adaNewb1234I honestly don't know where I would start. And given it all seems to be working...09:40
onefangadaNewb1234: "man service" might provide a clue. "The existence of a systemd unit of the same name as a script in /etc/init.d will cause the unit to take precedence over the init.d script."11:10
onefangOn my system there is indeed a /usr/lib/systemd/system/mariadb.service11:10
onefang"/etc/init.d/[whatever] directly" is what I do, or just let the sysvinit system start things on boot.11:11
adaNewb1234yes, but that doesn't really help with the post-install script.11:27
onefangIf the post install script is using /etc/init.d/mariadb and that's screwing up coz only the systemd unit works...11:36
onefangWhich version of Devuan?11:36
adaNewb1234a mix of chimaera and daedalus. I should probably fix that.11:43
onefangThat might help.11:47
adaNewb1234I had updates, security and backports for daedalus commented out. it's now updating.11:50
adaNewb1234it just tried to reconfigure mariadb-server. same problem.11:51
djphdaedalus has backports?12:04
adaNewb1234it's not complaining that it doesn't, so...12:06
onefangdaedalus-backports exists, but is unlikely to actually have anything in it yet.12:11
al1r4ddjph: i think yes14:56
alvhi all17:59
gnarfacehi alv, if you've got questions it's best to just ask them17:59
gnarfaceit can be a slow channel18:00
rwpIt's a support channel and questions take priority.  It can be very busy when solving a problem.18:15

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!