Message boards :
Questions and problems :
Why old virtual box comes with boinc download?
Message board moderation
Author | Message |
---|---|
Send message Joined: 28 Aug 19 Posts: 49 |
Ya, because if BOINC is going to a 3rd party program then BOINC should keep up to date with that 3rd party. |
Send message Joined: 29 Aug 05 Posts: 15483 |
The present BOINC available for download is 7.14.2 which was released in October 2018. The software release manager is not updating the installer every time VirtualBox or another 3rd party addition to BOINC gets a new update out, that'll be done once he's gearing for recommended release. Aside from that, in the past we've found that not all newest VBox versions were compatible with all projects out there. So sometimes it's better to use a trusted known to be working version than it is to update to the latest and greatest and find that it breaks projects for everyone using that version. |
Send message Joined: 5 Oct 06 Posts: 5082 |
I've posted a public note on the mailing lists about the need to choose a stable VBox version which meets the minimum needs of LHC for the next BOINC release, without necessarily being right on the bleeding edge. I've also suggested privately that BOINC and CERN (David and Laurence) should compare notes and choose a version. I'll keep an eye on how that goes. |
Send message Joined: 5 Oct 06 Posts: 5082 |
I'm in the process of sending an email to CERN, to ask that very question. |
Send message Joined: 28 Jun 10 Posts: 2539 |
Really?! I've never in my entire life had any new version of any software work worse than the older version. But I've lost count of how many times I've had unknown problems only to find there was a newer version to fix them that I didn't know about. That would be why almost all software now autoupdates. You have clearly not worked in the NHS. During my time there, more than once I found a new version of the software broke things. |
Send message Joined: 28 Jun 10 Posts: 2539 |
Why package VB with the Windows version at all? Surely Windows users are capable of finding it and installing it themselves just as Linux users do? |
Send message Joined: 5 Oct 06 Posts: 5082 |
It happened with the v7.16.1 release of BOINC, too. But at least that was intended as a test release (although it rather leaked out). We expected it to have problems - though perhaps not quite as spectacular as aborting the entire cache of WUs, completed, running and waiting.Really?! I've never in my entire life had any new version of any software work worse than the older version. But I've lost count of how many times I've had unknown problems only to find there was a newer version to fix them that I didn't know about. That would be why almost all software now autoupdates.You have clearly not worked in the NHS. During my time there, more than once I found a new version of the software broke things. |
Send message Joined: 5 Oct 06 Posts: 5082 |
I'm in the process of sending an email to CERN, to ask that very question.I've had a reply. CERN wrote: We don't have a strict dependency on the VBox version. The minimum version we accept is v5. The volunteers seem to have a good sense for what a good version is.Which doesn't take us much further forward. Can anybody here point me to one of those volunteers with good sense? Or a message board thread? |
Send message Joined: 5 Oct 06 Posts: 5082 |
OK, I'll stand down my earlier warning. But suggest that it's something worth considering - Jacob Klein is another bleeding-edger, who might have a view. |
Send message Joined: 5 Oct 06 Posts: 5082 |
From the opening post: I had to update it manually to run LHC properly.No hard feelings, but could all posters please be careful about the precise language they use? We do read these boards, and sometimes they're a very useful source of information about bugs or features that we need to pass on to the developers. In this case, it seems to have transpired that that phrase was, perhaps, a tad over the top? |
Send message Joined: 5 Oct 06 Posts: 5082 |
Don't worry - just touched a raw nerve. I've got an ongoing grumble with BOINC itself for sometimes emitting misleading error messages. |
Send message Joined: 29 Aug 05 Posts: 15483 |
From what I've read in LHC threads, there have never been problems relating to either version of virtualbox.Never say never, as a quick search reveals different. And remember that LHC isn't the only project using VirtualBox, so even if it may have sturdy software capable of running on (alpha) VBox versions without many problems, it doesn't mean that the other projects using VirtualBox can do that as well. You just need a handful of people crashing lots of models for a project to say "No!". But you were missing the point. The BOINC version available for download is from October 2018, not 2019. It has a 2018 version VBox 5.2 included, and while it may be an older version even for 2018 (February 2018), it's the version of choice by the release manager and the projects using VirtualBox. A user can always afterwards update to the latest greatest newest if they want to, but none of the projects require it. Aside from that, VBox 6.0+ is only for 64bit hosts, they no longer support 32bit hosts. A little thing BOINC and these projects still do. And lastly, there's the BOINC VboxWrapper. While it may run fine on some people's devices (hardware and OS) with the newest VBox, it may only produce garbage at others. The wrapper has to be developed in concert with the projects using VirtualBox and tested in a for them stable version. Which doesn't always mean the latest newest is best, no matter how many bugs it has fixed internally. |
Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License,
Version 1.2 or any later version published by the Free Software Foundation.