Showing headlines posted by dba477

Why packaging is important ? Original code was pretty straight forward Java coding which looks and works quite smoothly . However, all compiled classes appear to be placed in one folder WEB-INF/classes/ and in some way this minor design issue hides a business logic of apps.

As far as I can see any Java CRUD App would require a hack kind of described below , no matter is it "update" or "insert" or "select" JDBC setup doesn't work as it did earlier . Code bellow - invoking Connection "conn = DBConnection.createConnection();" along with localized JDBC setup in DBConnection.java works for me.

This post follows up "Practice coding in Java by writing a game" recently published at LXER. Focus has been done on manual invoking "javac" using standard command line options for placing in right folders packaged Java Beans and Servlets which are required by Tomcat 9 Java Web's server.

Even though snaps show was created in VENV, everything works the same way on bare metal. After you created the new partition of desired size installer brings you to screen which allows to redefine functionality of partition and assign name to folder supposed to be mounted on this partition . In case name of folder is not on the list you can create new folder's name manually . . . .

NVMe is not affected by the ATA interface constrictions as it sits right on the top of the PCI Express directly connected to the CPU. That results in 4 times faster Input/Output Operations Per Second (IOPs) rivaling the fastest SAS option out there. The seek time for data is ten times faster. NVMe can deliver sustained read-write speed of 2000MB per second, way faster than the SATA SSD III, which limits at 600MB per second. Here the bottleneck is NAND technology, which is rapidly advancing, which means we’ll likely see higher speeds soon with NVMe.

It appears that straight forward install Debian Testing via official ISO image at least in meantime hangs . Looks like "sddm" hangs attempting to pop up logging screen prompt . . .

At the time of writing KDE Plasma install on any SparkyLinux 2020.09 might be committed via GDM3 installation right after KDE Plasma due to after system reboot GDM seems to be the only one DM on Sparky detecting previously installed KDE.

This is a bit different method until Oracle release Fedora 33 repositories. Fedora 32 VirtualBox rpm works with Fedora 33, but unfortunately there is fixed version of python(abi) = 3.8, we have to change this and rebuild rpm package. Here is commands howto do it.

Due to Libvirt version 6.8.0 there is no more any problems with deployment KVM Guests on EXT4 file system.

Due to version LIBVIRT&&QEMU you are forced to deploy Ubuntu 20.10 on Btrfs file system . In particular case, Virtual Environment is a fair enough.

Attempt to test another Linux distro Manjaro KDE 20.1.1 as Virthost managed via Cockpit Web Console, i.e. deployment and management KVM guests on the recent Manjaro KDE utilizing Web Console . Bare metal platform is Ryzen 7 3700X based with 16 GB RAM and mother board MSI X570 A-PRO ( SVM status enabled )

In current post we intend to test Cocpit Web Console functionality on pre-release Fedora 33 Server based on the most recent Fedora nightly builds list . Start with deploying Server Fedora 33 as VM (L1) on Server Fedora 32 (bare metal)

I believe that this article is worth reading even in meantime as well as another one been written by the same author "QEMU Internals: Overall architecture and threading model"

Following below verifies Fedora 33 beta L1 Guest ability to perform KVM deployment Fedora 33 beta L2 Guest inside itself. Deployment is based on bridge to virtual interface of L1 F33 beta guest due to L1 guest has been deployed on external office network , L2 Guest also appears to be set up on the same external network.

Following below is a brief "howto" succeed with mentioned installation on SparkyLinux 2020.09 in meantime

Install Ubuntu 20.04.1 as Virtual Box Guest on Fedora 32 Server

First notice that we are installing on Ryzen box with board MSI X570 A-PRO in UEFI mode with "kvm" enabled. First of all we intend to verify that kvm has been enabled $ dmesg | grep kvm => [ 4.376968] kvm: Nested Virtualization enabled => [ 4.376978] SVM: kvm: Nested Paging enabled => and proceed with installing Virtualization Host the same way we would do it on Debian “bullseye”

First install "Development Tools", rebuild Python 3.8.5, after installation sqlite-devel. Afterwards build sqlitebrowser from source rather then install via snap, just a couple qt5-devel packages would have to be installed for successful build on CentOS 8.2.

Deployment of simple SparkyLinux2020.08 configuration on bare metal is supposed to be performed with "btrfs" file-systems for "/boot" and "/" .Otherwise virt-manager will report failure of deploying any guest due to /var/lib/libvirt/images/ "folder" , which for Libvirt 6.6.0 && Qemu 5.1.0, should already belong to "btrfs" formatted file-system at least at time of writing of this post.

The major target of posting bellow is supposed to demonstrate that virt-manager is not required to enable UEFI support for KVM Guest on Linux . See recent post at Lxer "Enable UEFI Support For KVM Virtual Machines In Linux". First - don't specify "Start immediately" and switch Firmware to "UEFI" been inside Web Cockpit Console