Bug 49940 - Check RAM over-commitment before live migration by only using available memory
Check RAM over-commitment before live migration by only using available memory
Status: CLOSED WONTFIX
Product: UCS
Classification: Unclassified
Component: Virtualization - UVMM
UCS 4.3
Other Linux
: P5 normal (vote)
: ---
Assigned To: UCS maintainers
UCS maintainers
:
Depends on: 48098
Blocks: 48901
  Show dependency treegraph
 
Reported: 2019-07-31 08:26 CEST by Christian Völker
Modified: 2023-06-28 10:46 CEST (History)
11 users (show)

See Also:
What kind of report is it?: Bug Report
What type of bug is this?: 7: Crash: Bug causes crash or data loss
Who will be affected by this bug?: 1: Will affect a very few installed domains
How will those affected feel about the bug?: 4: A User would return the product
User Pain: 0.160
Enterprise Customer affected?: Yes
School Customer affected?:
ISV affected?:
Waiting Support:
Flags outvoted (downgraded) after PO Review:
Ticket number: 2019073021000535
Bug group (optional): External feedback, Large environments, Usability
Max CVSS v3 score:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Christian Völker univentionstaff 2019-07-31 08:26:37 CEST
+++ This bug was initially created as a clone of Bug #48098 +++

UVMM should detect CPU/RAM/... over-commitment when a live migration is triggered and issue a warning.
A customer has regular problems with over-commitment because they have quite a lot of KVM servers and always checking resources of both migration partners quickly becomes a tedious task.


Additionally to the fix from the original bug we need to check for the migration only for available memory.
In customer environment after a live migration the host (and all VMs) went unusable due to the high amount of swap in/out or low memory ressources.

The check should consider the possiblity of a memory leak and therefor not use the total amount of memory but the available amount of memory (which is free+bug/cache).
Comment 1 Oliver Bohlen 2019-07-31 09:47:52 CEST
I think the best/savest solution should be:

Free/Useable RAM for new/migrated VMs = "Total RAM by e.g. by free" - "Max useable RAM of ALL VMs" - "RAM used by Host System excl. VMs" - "Buffer from uvmm/overcommit/reserved e.g for system qemu-Processes"
Comment 2 Ingo Steuwer univentionstaff 2019-08-13 08:41:08 CEST
We should add an additional check to test whether the configured RAM of a to be migrated or to be started virtual machine fits in the currently available free memory.

We can not check whether some process might have a memory leak now or in the future. This needs to be monitored with other available tools (Nagios, UCS Dashboard) as it can occur at any time, not only while a virtual machine is started / migrated.

The "uvmm/overcommit/reserved" configuration is meant to include both the needed RAM for standard services on the node and RAM needed to manage virtual machines and other QEMU related needs.
Comment 3 Ingo Steuwer univentionstaff 2020-04-17 14:32:21 CEST
From my understanding the support case related to this Bug Report is closed, so I unset the "Waiting for Support" flag.
Comment 4 Ingo Steuwer univentionstaff 2021-05-14 13:46:11 CEST
This issue has been filed against UCS 4.3.

UCS 4.3 is out of maintenance and many UCS components have changed in later releases. Thus, this issue is now being closed.

If this issue still occurs in newer UCS versions, please use "Clone this bug" or reopen it and update the UCS version. In this case please provide detailed information on how this issue is affecting you.