RBVM Bugs: Difference between revisions
Jump to navigation
Jump to search
No edit summary |
No edit summary |
||
| Line 6: | Line 6: | ||
## Addressed a little in rev 70, still not quite there. | ## Addressed a little in rev 70, still not quite there. | ||
## This is happening at random again. Changed this from improvement to major bug, and added some debug info to figure out why this still isn't working 100% in r76. | ## This is happening at random again. Changed this from improvement to major bug, and added some debug info to figure out why this still isn't working 100% in r76. | ||
## This seems to have stopped happening since power on detection rules were relaxed (/proc timestamp checking was removed). Keep an eye on it though. | |||
# <span style="color:#fff;background:#f00">Unassigned</span> '''Moderate Bug:''' Session cache should be cleared when the system starts up, otherwise lock errors cause some peoples' sessions to asplode. | # <span style="color:#fff;background:#f00">Unassigned</span> '''Moderate Bug:''' Session cache should be cleared when the system starts up, otherwise lock errors cause some peoples' sessions to asplode. | ||
# <span style="color:#fff;background:#f00">Unassigned</span> '''Improvement:''' Logging and debug levels, etc, need to be cleaned up properly. | # <span style="color:#fff;background:#f00">Unassigned</span> '''Improvement:''' Logging and debug levels, etc, need to be cleaned up properly. | ||
| Line 14: | Line 15: | ||
# <span style="color:#fff;background:#080">Fixed: r73</span> '''Minor Bug:''' Live boot device selection didn't interact with KVM properly | # <span style="color:#fff;background:#080">Fixed: r73</span> '''Minor Bug:''' Live boot device selection didn't interact with KVM properly | ||
# <span style="color:#fff;background:#080">Fixed: r75</span> '''Improvement:''' Add ACPI control (-no-acpi KVM cmd line option) to KVM module and frontent. | # <span style="color:#fff;background:#080">Fixed: r75</span> '''Improvement:''' Add ACPI control (-no-acpi KVM cmd line option) to KVM module and frontent. | ||
# <span style="color:#fff;background:# | # <span style="color:#fff;background:#080">Invalid</span> '''Moderate Bug:''' IO performance sucks since a kernel update last week. Some security patches for KVM were involved. Investigate. | ||
## This turned out to be a misconfiguration on daniel. Not a bug in the RBVM source code. | |||
{{rbvm}} | {{rbvm}} | ||
__NOTOC__ | __NOTOC__ | ||
[[Category: RedBrick VM Project]] | [[Category: RedBrick VM Project]] | ||
Revision as of 22:05, 1 May 2010
This page lists bugs with the redbrick VM system (at https://vmweb.redbrick.dcu.ie, documentation here).
- Unassigned Minor Bug: VM names can be set to empty string. Cosmetic bug only.
- Fixed: r70 Major Bug: For some reason, on VM power on, the os.read() call on the monitor pseudoterminal blocks occasionally, causing the power on to fail (and the KVM process to stay running) (removed use of PTYs)
- In progress (werdz) Major Bug: VM powered on detection needs work, it should check by MAC address and disk image name to try to detect when a VM power on failure has occurred.
- Addressed a little in rev 70, still not quite there.
- This is happening at random again. Changed this from improvement to major bug, and added some debug info to figure out why this still isn't working 100% in r76.
- This seems to have stopped happening since power on detection rules were relaxed (/proc timestamp checking was removed). Keep an eye on it though.
- Unassigned Moderate Bug: Session cache should be cleared when the system starts up, otherwise lock errors cause some peoples' sessions to asplode.
- Unassigned Improvement: Logging and debug levels, etc, need to be cleaned up properly.
- Fixed: r66 Moderate Bug: If new VMs are created as root, the mode and ownerships of the VM images under /var/lib/rbvm/images is completely wrong - VM will refuse to start as the rbvm user can't read/write the image.
- Fixed: r65 Improvement: Add extra columns (MAC, IP, status) to homepage.
- Fixed: r72 Improvement: Add hardware options to front end (NIC controller, etc). Backend added in rev 70
- Fixed: r69 Improvement: Boot device selection
- Fixed: r73 Minor Bug: Live boot device selection didn't interact with KVM properly
- Fixed: r75 Improvement: Add ACPI control (-no-acpi KVM cmd line option) to KVM module and frontent.
- Invalid Moderate Bug: IO performance sucks since a kernel update last week. Some security patches for KVM were involved. Investigate.
- This turned out to be a misconfiguration on daniel. Not a bug in the RBVM source code.