=== Napsterbater is now known as Guest7978 | ||
=== Napsterbater_ is now known as Napsterbater | ||
=== jelly-home is now known as jelly | ||
=== Napsterbater is now known as Guest81968 | ||
=== Napsterbater_ is now known as Napsterbater | ||
=== Napsterbater is now known as Guest86747 | ||
=== Napsterbater_ is now known as Napsterbater | ||
=== Napsterbater is now known as Guest5102 | ||
=== Napsterbater_ is now known as Napsterbater | ||
de-facto | where is the common network config ? | 08:46 |
---|---|---|
Slashman | hello, on Ubuntu 20.04, why is there 'GRUB_CMDLINE_LINUX_DEFAULT="maybe-ubiquity"' in /etc/default/grub after a server install? is it safe to remove it? | 09:15 |
Slashman | I also see a process "/sbin/init maybe-ubiquity" | 09:15 |
maret | Hi, I would like to figure out what was the highest values per process (run in screen) or just per last 24 hours for CPU, IO and Memory. Simple I am running few processes in paralel over night and I need to find out whats bottleneck | 10:25 |
=== StathisA_ is now known as StathisA | ||
hydrian | Ello all | 17:07 |
=== ijohnson is now known as ijohnson|lunch | ||
hydrian | I have LVM VG Storage for my KVM instance. I expanded one of the LVs fot one of the disks that active in one of my guest VMs. | 17:10 |
hydrian | KVM doesn't see that the LV has been expanded (via lvextend). | 17:11 |
hydrian | Is there a way to force a refresh without restarting the system? | 17:12 |
hydrian | After a `virsh pool-refresh --pool POOL_NAME` the KVM pool sees the difference. | 17:21 |
hydrian | Now how do I get the guest VM to see the change? I tried rescanning the SCSI drive bus but it isn't helping. | 17:22 |
hydrian | This is a virt-block device. | 17:24 |
=== ijohnson|lunch is now known as ijohnson | ||
RoyK | hydrian: no idea - but perhaps the guys at #virt @ irc.oftc.net can help | 18:30 |
tomreyn | hydrian: hmm, i guess that's what i'd done as well, maybe run partprobe, too. In the end, since this will likely be a partitioned device inside the guest, you'll need to repartition to span the full disk, then resize file systems (or intermediary block device layers, if any) | 19:35 |
hydrian | tomreyn: The problem was that the Guest OS didn't get notification to update the geometry of the drive. | 19:55 |
hydrian | Typically existing drives don't change geometry once installed. They can be added/removed but don't change. Partitions change though. | 19:57 |
tomreyn | hydrian: but the scsi bus scan should have helped there, shouldn't it? i guess i doesn't work in all scenarios. | 20:18 |
hydrian | It didn't. Not sure if it is because virtio-block runs on the 'ATA' bus. I though all of the ATA devices have been backed by the SCSI bus a long time ago in the kernel. Like 2.6 days. | 20:23 |
sdeziel | hydrian: https://www.ocf.berkeley.edu/docs/staff/procedures/live-resize/ seems to suggest virtio-blk support live resize | 20:23 |
sdeziel | hydrian: can you check in dmesg? | 20:24 |
hydrian | sdeziel: I was | 20:25 |
hydrian | When I'd rescan the SCSI bus, I'd see the bus scan, but no changes. | 20:26 |
sdeziel | I'd guess that `virsh blockresize` signals qemu about the changes to the underlying disk | 20:26 |
hydrian | To verify, I would run parted /dev/vdd and check the size | 20:26 |
hydrian | sdeziel: I couldn't get that to work. | 20:26 |
hydrian | It may because I'm not using a file for VM. I'm using RAW LVM. | 20:27 |
hydrian | The working examples was with an .IMG file. | 20:28 |
sdeziel | lvm is used in the link I provide | 20:28 |
sdeziel | hydrian: what error did you get from blockresize? | 20:29 |
hydrian | virsh # qemu-monitor-command Ruadan block_resize --hmp drive-virtio-disk8 90G | 20:30 |
hydrian | Error: Cannot find device=drive-virtio-disk8 nor node_name= | 20:30 |
hydrian | Ruadan is the VM. I verifed the device name many times overy. | 20:31 |
hydrian | over | 20:31 |
hydrian | sdeziel: ^ | 20:32 |
sdeziel | hydrian: I don't know about qemu-monitor-command but I'd give "sudo virsh blockresize $VM /dev/vg/$VM <size in bytes>B" a try | 20:34 |
hydrian | sdeziel: Didn't know that command existed. All of info I found from google said I had to go through the qemu-monitor command or it wasn't possible. Most of it was from the RHEL 5-6 days though. | 20:36 |
hydrian | I gave up and bounced the machine a while back. | 20:36 |
sdeziel | oh OK :) | 20:36 |
hydrian | I am trying to update my wiki for how I do this. I always forget how to do but I end up doing once every 6 month and forget how to do it every time. | 20:37 |
sdeziel | VMs are cheap, you can always create a dummy VM to document the process ;) | 20:45 |
hydrian | Yea... but this is just personal. Not work situation. | 20:47 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!