| lordievader | Good morning | 06:30 |
|---|---|---|
| uzee | Hi All, does anyone know how lvm partitioning can be done in preseed but based on percentage of disk space, instead of absolute numbers for min/max/priority ? | 06:53 |
| _KaszpiR_ | d-i partman-auto-lvm/guided_size string 90% | 07:18 |
| _KaszpiR_ | or just create simple partitions and use script to make your own customs | 07:21 |
| chl_ | anyone know where static leases are written to in isc-dhcp, when you add them by omapi? | 07:30 |
| uzee | _KaszpiR: thanks but "d-i partman-auto-lvm/guided_size string 90%" would give me how much a logical volume can use from the volume group, no? | 07:42 |
| uzee | sorry, I meant how much of the volume group to use | 07:42 |
| uzee | _KaszpiR: Also, not sure what you mean by creating simple partitions and then use script, do you mean create non-lvm partitions and then use like a preseed/late_command or something? | 07:44 |
| uzee | The first thing I'm trying to understand is that if its even possible to use percentage values in the min, priority and max fields | 07:45 |
| uzee | It seems counter-productive to allow absolute values only, as that implies knowing my disk sizes everytime before hand. While thats doable, but we have an almost 100% virtual env. and sometimes VM creations from templates or clones is carried out without giving a lot of thought to storage size | 07:48 |
| uzee | IMHO percentages would allow to do the right thing whether a VM is created with a 15GB disk or 150GB | 07:49 |
| weedmic | cheers | 08:52 |
| Onepamopa | Q: what's the best way to transition ubuntu xenial 16.04.4 from openssl 1.0 to 1.1 ? | 09:57 |
| Onepamopa | (system-wide) | 09:57 |
| Onepamopa | if that's even possible | 09:57 |
| rbasak | Not really practical. | 10:04 |
| rbasak | 18.04 ships 1.1. You could put a 18.04 container on a 16.04 system if you can't upgrade it. | 10:06 |
| Onepamopa | well, I could but then I'd have to fix 50 custom software-related s***s ... | 10:07 |
| Onepamopa | + directadmin is running there ... it'd be a mess | 10:07 |
| supaman | trying to set up an ldap server which is working, I can connect to it using basic slapd tools, but when I try to connect to it using tools from ldapscripts like ldapadduser then I get a error 49 (permission denied or wrong password), I set the debug level to -1 and here is the result from one ldapadduser try: http://paste.debian.net/1081705/ | 10:31 |
| supaman | can someone here see what is wrong? | 10:31 |
| supaman | I have double checked the password in /etc/ldapscripts/ldapscripts.passwd and it does not contain a trailing newline | 10:35 |
| supaman | dammit, the password is wrong | 10:51 |
| uzee | Asking again in the hopes that some more folks would've come online and I might be able to get clarity | 11:06 |
| uzee | I'm trying to do lvm partitioning in a preseed file and can't figure out if I can use percentages for min, priority and max fields..? anyone? | 11:07 |
| uzee | Absolute values work fine but I'd like to setup a preseed where I don't have to keep changing the values based on the disk size of each server I provision | 11:07 |
| uzee | hence percentages would be much cleaner solution, can anyone advise if thats possible? I've tried sticking in percentage values but the install fails at a later time with "no space left" error | 11:08 |
| rbasak | With partman? | 11:19 |
| rbasak | I've tried before and never got anything useful out of it. The algorithm is well specified, but it cannot be easily reversed, IIRC. | 11:20 |
| tomreyn | uzee: since lvm is so flexible, you could just create those LVs with small (but large enough to install ubuntu) fixed sizes, then resize them according to your needs post install. | 11:30 |
| lordievader | I'd second that approach. LVM with ext4 can easily grown. Shrinking is a bit harder. So determine what you need. Add a bit of headroom and use those values in the preseed. | 11:32 |
| gislaved | lordievader why shrink ? never reducre your private collection ;) | 12:21 |
| gislaved | *reduce | 12:21 |
| uzee | thanks much rbasak, tomreyn and lordievader | 12:21 |
| uzee | My fundamental thing was to first verify if partman can actually handle percentage values. Looks like it can't. You'll are right, I also thought of the same approach to create small enough lvm partitions and then resize later. I use ansible for config mgmt, so that can be automated as well, but again, for potentially every different disk-sized server, I will need to individually address it at either preseed or ansible :( | 12:23 |
| uzee | kickstart on the other hand handles it without any issues. I could be wrong but my thinking is that a percentage approach allows to have a single kickstart file for me regardless of disk sizes, the partitions will always be proportional | 12:25 |
| samba35 | i am trying to setup pci passtrough with 16.04.06 i am able to start guest with no error but i could not see another guest on second monitor | 14:45 |
| samba35 | i have blacklisted nvida for guest /monitor | 14:57 |
| samba35 | be right back | 15:11 |
| bobbytables5_ | hello, where should a cloud-config yaml should be put ? or if I want to run a bash script on first launch, is it enough to copy it somewhere for cloud-init to start it ? | 16:16 |
| blackboxsw | bobbytables5_: if you are rolling your own images, you can place cloud-config yaml in a file under /etc/cloud/cloud.cfg.d/ in that image. you can use bootcmd: config directive in that case per https://cloudinit.readthedocs.io/en/latest/topics/modules.html#bootcmd | 16:18 |
| blackboxsw | bobbytables5_: you can also provide that as user-data to most clouds when launching an instance either through their web UI or CLI tools. cloud-config userdata just needs to be prefixed with #cloud-config on the first line. here are some examples https://cloudinit.readthedocs.io/en/latest/topics/examples.html#yaml-examples | 16:20 |
| blackboxsw | bobbytables5_: alternative number 3: put your executable script in /var/lib/cloud/scripts/per-instance/ or per-once/ | 16:22 |
| bobbytables5_ | blackboxsw: had to delete my response because I think alternative number 3 would be the perfect solution | 16:23 |
| blackboxsw | sounds good :) | 16:23 |
| bobbytables5_ | blackboxsw: /var/lib/cloud/scripts/per-once/myscript.sh just putting it here should be enough ? I created an AMI and a new instance but in the scripts there is trace of it beeing called :( | 16:39 |
| bobbytables5_ | no trace* | 16:39 |
| blackboxsw | bobbytables5_: per-once is only called on first clean boot of cloud-init ...... if you have cloud-init version 18.3 or later you can run sudo cloud-init --logs --reboot.. then check /var/log/cloud-init.log for logs related to per_boot | 16:41 |
| blackboxsw | sorry I mean per_once not per_boot | 16:42 |
| bobbytables5_ | ah I see, it should have been on per-instance | 16:43 |
| blackboxsw | per-instance would work as you end up booting a new instance id from your snapshot AMI | 16:43 |
| blackboxsw | right | 16:43 |
| blackboxsw | output or errors from your script will typically end up in /var/log/cloud-init-output.log FYI | 16:44 |
| ninekeys | Anyone know how I can get ahold of the racadm util? Dell's site has stuff for RHEL but their Debian/Ubuntu sections are lacking/broken | 19:02 |
| tomreyn | http://linux.dell.com/repo/community/openmanage/ | 19:06 |
| ninekeys | tomreyn: Thanks! I'll give that a shot! | 19:07 |
| tomreyn | you're welcome | 19:09 |
| RoyK | dell is usually very aqueinted to redhat and not with open systems like debian | 19:28 |
| ninekeys | RoyK: Yea, damn shame too. That's one of the reasons why most of the boxes here are CentOS. | 19:54 |
| === Sveta is now known as Guest31756 | ||
| === svetlana is now known as Sveta | ||
| RoyK | ninekeys: same reason why my boxes run debian | 22:41 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!