Werner Fischer in Austria has done development work with OpenVZ and high availability clustering, which he says, “makes it possible to start a virtual machine in seconds after a failover” within the information technology infrastructure. Werner Fischer is a developer at Thomas-Krenn.AG. He recently presented a paper on the subject at the Linux Tag conference in Germany (on May 6).
As might be expected, the checkpointing code is on the long and complicated side. The checkpoint process starts by putting the target process(es) on hold, in a manner similar to what the software suspend code does. Then it comes down to a long series of routines which serialize and write out every data structure and bit of memory associated with a virtual environment. The obvious things are saved: process memory, open files, etc. But the code must also save the full state of each TCP socket (including the backlog of sk_buff structures waiting to be processed), connection tracking information, signal handling status, SYSV IPC information, file descriptors obtained via Unix-domain sockets, asynchronous I/O operations, memory mappings, filesystem namespaces, data in tmpfs files, tty settings, file locks, epoll() file descriptors, accounting information, and more.
Cedrik Le Goater of IBM France just wrote a mail to LKML and openvz-devel lists saying that IBM is testing different virtualization projects. Here is what he wrote:
Recently, we've been running tests and benchmarks in different virtualization environments : openvz, vserver, vserver in a minimal context and also Xen as a reference in the virtual machine world.
We ran the usual benchmarks, dbench, tbench, lmbench, kernerl build, on the native kernel, on the patched kernel and in each virtualized environment. ( Read more...Collapse )
I am personally very happy to get this news. Let me explain why.
We do a lot of tests for OpenVZ, including tests for performance and stability, trying hard to find and fix all possible problems before they occur on users' boxes. We also do comparisons with the other projects — like Xen, UML or Linux-VServer — to make sure we are at least on a par.
It doesn't make much sense for OpenVZ to prepare and publish benchmarks. A comparison should be done by a trusted, independed third party, which will perform in a fair and unbiased manner. By its own initiative, IBM is fulfilling a meaningful role here.
Clearly, this work benefits users of virtualization technology. It would be terrific, I think, if the whole process is open in a collaborative community effort and all the projects are involved as active participants.
I am looking forward to seeing the test results, and contributing to the testing scenarios, methodologies and cases.
Do you still stand by your opinions above now in 2016?…