[03:23] hey all, i have ubuntu server running on what is essentially an old desktop system. i have emby set up in a docker container. it had been streaming smoothly to even a couple devices at a time in the house. but lately devices have been buffering occasionally. the buffering devices are all wireless, three roku units and a laptop. the server is connected to a switch along with two other computers, which is then connected to the network extender [03:23] unit, connected to the main modem at the other end of the house. [03:24] is there a way to see if something in my network is causing the buffering, such as the switch, or a wireless router? [13:25] rharper: when you get in. growpart .. your fix for bug 1834875 [13:25] bug 1834875 in cloud-utils (Ubuntu Eoan) "cloud-init growpart race with udev" [Undecided,Fix committed] https://launchpad.net/bugs/1834875 [13:25] i dont think that unlock_disk_and_settle ever gets called. [13:26] because $resizer (resize_sfdisk_gpt, resize_sfdisk_dos, resize_sgdisk_gpt) does not return. [13:53] hey jamespage - could you do a follow up review on https://code.launchpad.net/~chris.macnaughton/ubuntu/+source/python-sqlalchemy-utils/+git/python-sqlalchemy-utils/+merge/386651 again? I'd like to get +1 from you and coreycb [14:18] anybody experienced with zfs zvol? I created a virtual machine using a zvol as storage. The virtual machines is using ext4 as Filesystem. Do I have to take care of the block size? volblocksize (zfs) is 8K, ext4 block size is 4K. Is that a problem? [14:20] Aison0: I use zvols as backing store for VMs almost exclusively. I've not had issues where I felt pressed to muck with block sizes thus far. [14:21] mason, ok [14:22] mason, one strange thing: I allocated a zvol with a size of 25G. The guest vm can use this 25G. Why does the zvol use over 50G? [14:22] icey: looking now [14:22] Aison0: It might make more of a difference if your back end is flash. [14:22] logicalused 25.1G, logicalreferenced 25.1G [14:22] Aison0: Snapshots maybe? [14:22] Aison0: ZFS space usage is kind of a black art. [14:23] used 54.1G, referenced 53.5G [14:23] thanks jamespage [14:23] there are no snapshots [14:23] Aison0: #zfsonlinux has folks way more expert than I that can probably give deeper, better answers. [14:23] mason, :-) thx for that hint [14:24] Aison0: From what I see, my backing zvols tend to be right around the size I allocated, with refer slowly growing as the virtual disks fill. [15:11] paride: i guess cloud-utils got moved to github ? [15:12] smoser, yes [15:12] i think probalby you should just remove the launchpad git repo [15:12] unless you have some reason for it [15:20] jamespage: I've updated with your suggestion too now :) [15:21] smoser, we discussed the thing a bit. We'll keep using the LP bug tracker as it has per-series bugs and allows closing bug from d/changelog [15:21] we could do this without the repo, but the repo allows to have daily builds in PPAs [15:22] and LP is setup to automatically mirror GitHub [15:23] so I think we're keeping it as it comes almost for free and allows recipies [15:23] *recipes [15:24] smoser, I'd like to disable MPs to the repo on LP, but I don't think that's possible [16:40] Is there any way to prevent the 20.04 -server installer, when creating a LVM VG, from zeroing it? [16:44] is't this like a tickbox option when you configure it? [16:45] i guess the fact that you'Re asking this contains the answer to my question. ;-) [17:22] indeed; I had already created an underlying MD RAID1 from shell using mdadm ... --assume-clean ... to avoid a long sync and had manually created a VG on it. The installer could not refresh the current view to pick up the VG and on rebooting the installer it wouldn't offer the existing VG as an installation target, so had to use the option to create a VG which does the zeroing [17:27] right, i don't think it's particularly good at handling existing structres. === coconut_ is now known as coconut