|
I started over and today UEFI seems to work. Annoying to find out however Windows 10 setup doesn't have floppy support any more to install storage drivers.
|
#
?
Apr 3, 2016 10:50
|
|
|
|
| # ? May 8, 2024 12:01 |
|
|
Riso posted:I started over and today UEFI seems to work. What the heck? Are you trying to install a Bernoulli drive or something?
|
|
#
?
Apr 3, 2016 14:48
|
|
|
Just using qemu I have passed through the gpu with vfio and I have a picture! Of Windows blue screening on start. BAD_SYSTEM_CONFIG?? Alright, let's pack this in and try again some other time. HPL posted:What the heck? Are you trying to install a Bernoulli drive or something? Trying to do Virtio storage drivers. Needless to say, didn't work. I remembered to attach the virtio iso as second cd drive but Windows setup just gave me an error at the end anyway. Riso fucked around with this message at 15:00 on Apr 3, 2016 |
|
#
?
Apr 3, 2016 14:56
|
|
|
Install with the emulated IDE adapter, then change the drivers of said adapter to the virtio ones. They'll only apply after rebooting, so shutdown the VM and change the stuff in QEMU to virtio.
|
|
#
?
Apr 3, 2016 15:24
|
|
|
I tried that. Works with Win8 but Windows 10 just decides to poo poo the bed even though it has drivers and everything.
|
|
#
?
Apr 3, 2016 15:39
|
|
|
Windows setup won't even detect the storage without virtio drivers. Boot with -drive file=...,if=virtio Also, please start providing information people may need, or move these updates to twitter/your blog. You got BAD_SYSTEM_CONFIG_INFO? And you had video? Why not a picture? And we have no idea whether that's qxl, whether that's on the passthrough GPU (which you should add after) Windows is installed/working, to make troubleshooting problems easier). We still don't know whether you used vfio, or what your qemu cmdline is, etc. It's kind of useless. And I'm not really trying to be bitchy, but it's frustrating. I get enough "30% of the info you need" questions on IRC.
|
|
#
?
Apr 3, 2016 16:09
|
|
|
I've installed virtio drivers with Windows 10 tons of times. On the page where it asks you which drive you want to install Windows on, there's a thing you click to load additional drivers. You do that, browse to the virtio driver ISO and then load up your drivers from there. This way, you can set up your system to use as many virtio devices as possible right off the bat.
|
|
#
?
Apr 3, 2016 16:51
|
|
|
Right. It's very easy. I just suspect from the description that it was booted/installed with a different bus, and not virtio at all. If virtio is used, Windows won't show any disks until the drivers are added from the iso
|
|
#
?
Apr 3, 2016 19:42
|
|
|
evol262 posted:Right. It's very easy. I just suspect from the description that it was booted/installed with a different bus, and not virtio at all. If virtio is used, Windows won't show any disks until the drivers are added from the iso After you add the virtio-0.1.102 (stable) drivers, the disk shows but trying to install setup throws error 0x80300002, can't install on media. Turns out that's a bug in the virtio drivers and it works fine with 0.1.113 (latest). Btw trying to add the qxl drivers in setup hard-locks the VM so I wouldn't recommend that. Everything seems to work fine now; the latest nvidia whqls didn't complain and installed fine. I now just have to work out how to share keyboard and mouse because I don't really want to give up control or handle several devices. Besides games I'll probably try remote desktop or synergy. linux mint 17.3, custom 4.4.6 kernel, qemu 2.1.2, OVMF 20160104 qemu-system-x86_64 -enable-kvm -cpu host,kvm=off -m 4G -smp cores=2,threads=1 -vga none -device vfio-pci,host=01:00.0 -device vfio-pci,host=01:00.1 -drive if=pflash,format=raw,readonly,file=/usr/share/OVMF/OVMF_CODE.fd -drive if=pflash,format=raw,file=win10_vars.fd -drive if=virtio,file=win10.qcow2 -usb -usbdevice host:046d:c069 -netdev user,id=network0 -device virtio-net-pci,netdev=network0 Riso fucked around with this message at 21:14 on Apr 3, 2016 |
|
#
?
Apr 3, 2016 21:02
|
|
|
qemu-git has a new input device object, with which you can switch the keyboard (--edit: and mouse) between guest and host by tapping both control keys. Also, why are you running a relatively new kernel and stick with qemu 2.1? Combat Pretzel fucked around with this message at 21:18 on Apr 3, 2016 |
|
#
?
Apr 3, 2016 21:12
|
|
|
Combat Pretzel posted:qemu-git has a new input device object, with which you can switch the keyboard between guest and host by tapping both control keys. I figured going back to something that ought to work for 14.04 eliminates a few problems.
|
|
#
?
Apr 3, 2016 21:17
|
|
|
I've made tons of Windows VMs from scratch with full virtio devices and the installs have gone fine. Mind you, I'm using virt-manager, not straight command line. So maybe make a VM with virt-manager and then check the resulting config file.
|
|
#
?
Apr 3, 2016 21:36
|
|
|
Synergy, or just use steam streaming from the vm
|
|
#
?
Apr 3, 2016 22:02
|
|
|
Tried synergy and it works pretty nice except the elevated security desktop tells it to gently caress right off. I'll try steam streaming. Other question. How do I pass my sata disk on /dev/sdb1 to qemu? Preferably using virtio and read-only. Documentation is all over the place, out of date or not helpful. Is "-drive if=virtio,readonly,cache=none,file=/dev/sdb"1 what I want?
|
|
#
?
Apr 4, 2016 10:17
|
|
|
You need format=raw, and you probably don't want readonly. But don't do this. Windows will see the underlying hardware change every time you reboot, and try reactivation.
|
|
#
?
Apr 4, 2016 13:46
|
|
|
evol262 posted:You need format=raw, and you probably don't want readonly. It will reactivate from passing it different disks?
|
|
#
?
Apr 4, 2016 14:43
|
|
|
cliffy posted:It will reactivate from passing it different disks? Passing in a raw disk with a Windows installation will work. But Windows will basically see an entirely new system (new DMI, etc), and reactivate. Then do it again when you boot back into Windows on your physical hardware. Rinse and repeat, until you're out of automatic activations.
|
|
#
?
Apr 4, 2016 15:10
|
|
|
evol262 posted:You need format=raw, and you probably don't want readonly. Thanks for the warning, but it's only my data drive so Windows shouldn't go nuts. Also I wanted it read only to make sure it doesn't gently caress anything up.
|
|
#
?
Apr 4, 2016 18:32
|
|
|
KillHour posted:I'll bite. Why? Why not? It was the 90's, dumb things happened.
|
|
#
?
Apr 4, 2016 20:52
|
|
|
The real reason is that someone decided that you should be able to resolve any share name through NetBIOS to any building on the MAN without needing to configure WINS or DNS, so the network got built around that.
|
|
#
?
Apr 4, 2016 20:54
|
|
|
|
|
#
?
Apr 4, 2016 20:56
|
|
|
BangersInMyKnickers posted:The real reason is that someone decided that you should be able to resolve any share name through NetBIOS to any building on the MAN without needing to configure WINS or DNS, so the network got built around that. There it is.
|
|
#
?
Apr 4, 2016 20:57
|
|
|
The more time I spend working on other people's networks the less surprised I am when the reason for a dumb setup comes down to "previous admin/decision maker doesn't understand/trust/care to properly implement DNS".
|
|
#
?
Apr 4, 2016 23:46
|
|
|
I work with people who don't 'trust' DNS or static DHCP entries so everything is hard coded to use IP addresses. You might think this makes it tough to have valid certificates, and you'd be right. The solution is to tell their clients to ignore SSL errors.
|
|
#
?
Apr 4, 2016 23:58
|
|
|
wolrah posted:The more time I spend working on other people's networks the less surprised I am when the reason for a dumb setup comes down to "previous admin/decision maker doesn't understand/trust/care to properly implement DNS".
|
|
#
?
Apr 5, 2016 00:00
|
|
|
AD has been supporting secure record updates since at least 2003, maybe even 2000. The excuses ended there.
|
|
#
?
Apr 5, 2016 00:01
|
|
|
BangersInMyKnickers posted:AD has been supporting secure record updates since at least 2003, maybe even 2000. The excuses ended there.
|
|
#
?
Apr 5, 2016 00:10
|
|
|
When I was doing a DC upgrade I wrongly assumed that because all the client machines were using DHCP for IP addressing they must also be using it to pull their DNS. So when I changed the DNS server scope option to what would be the new DC nobody except for my own computer and the 6 or so machines that I had personally set up got the message to use the new DNS server. 
|
|
#
?
Apr 5, 2016 00:36
|
|
|
adorai posted:To be fair, we had the occasional hosed up problem with DNS in 2003 and 2003r2. I'm not saying they weren't attributable to us, but we haven't had the same sort of problems in newer versions.
|
|
#
?
Apr 5, 2016 01:00
|
|
|
Vulture Culture posted:I don't think scavenging ever worked close to correctly until 2008 either I had zero problems with it on 2003 DCs and they didn't let me upgrade to 2012R2 until last year. It was fine. The biggest problem is client-side where Windows systems will delete their A record when they enter sleep though I think they finally patched that out.
|
|
#
?
Apr 5, 2016 01:54
|
|
|
god dammit boss. Currently sitting at 99% after an hour and a half removing the snapshot, I wonder how long it will take.
|
|
#
?
Apr 6, 2016 20:34
|
|
|
Did they take a snapshot at initial server build and let the entire volume be backed by a delta disk?
|
|
#
?
Apr 6, 2016 21:19
|
|
|
We have a lovely erp system that does a lot of unnecessary data churn nightly, which makes backups...unpleasant. Boss just forgot that he took the snapshot prior to a schema update and here we are. I really need to get a powercli script set up that will send a warning email for any snapshots >3 days old.
|
|
#
?
Apr 6, 2016 22:28
|
|
|
devmd01 posted:Currently sitting at 99% after an hour and a half removing the snapshot, I wonder how long it will take. We've had some large ones at work ~1TB that was roughly 3 days at 99% to remove due to ongoing IO load. Might be ~24hrs depending on your infrastructure.
|
|
#
?
Apr 7, 2016 04:24
|
|
|
Yeah, I have had snapshots take days to remove. I think there is a built in alert for old snapshots, but maybe I'm misremembering?
|
|
#
?
Apr 7, 2016 05:14
|
|
|
Well it got worse last night because my boss was impatient and start messing with it after hours. I was gonna give it the night but he found some guide on how to fix the issue with the stuck snapshot, meanwhile the host that the guest is on has a management agent issue. I start seeing vsphere alerts and hop on to take a look. After investigating I took a guess that the two problems were related, vmotion everything off of that host except for the messed up VM, and hard reboot the host since it wouldn't go into maintenance mode. HA took over and restarted that guest. When that host came back up and rejoined the cluster though, all hell broke loose with the ha agents on every host. They started fighting each other and trying to force HA failovers for every guest in the cluster which thankfully failed, and as I'm watching this all the guests are going from green to disconnected back to green every few seconds like a chasing christmas light. As i'm making GBS threads bricks for a couple of minutes making sure everything is still up (which it was, nothing got dropped), the cluster finally converges and settles down. I initiate the problem guest consolidation and go to bed, it's at 37% this morning. God I hate iSCSI. *drinks*
|
|
#
?
Apr 7, 2016 12:30
|
|
|
devmd01 posted:God I hate iSCSI. *drinks*
|
|
#
?
Apr 7, 2016 13:56
|
|
|
Sometimes you just read posts and cringe. That was one of those posts.
|
|
#
?
Apr 7, 2016 14:10
|
|
|
Internet Explorer posted:Sometimes you just read posts and cringe. That was one of those posts.
|
|
#
?
Apr 7, 2016 14:22
|
|
|
|
| # ? May 8, 2024 12:01 |
|
|
adorai posted:How is someone creating a snapshot and forgetting about it, letting it grow to gargantuan proportions an iSCSI problem? It's not, but the slowness of our current san makes that process more painful with the additional I/O. I've always been much happier with fiber channel. Brand new SAN with SSD tier in July, wooo!
|
|
#
?
Apr 7, 2016 14:27
|
|