Note: This is a beta release of Red Hat Bugzilla 5.0. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Also email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback here.
Bug 1355835 - [RFE] - Clarify possible overflow of total RX/TX byte statistics for hypervisor and VM network interfaces
Summary: [RFE] - Clarify possible overflow of total RX/TX byte statistics for hypervis...
Keywords:
Status: NEW
Alias: None
Product: Red Hat Enterprise Virtualization Manager
Classification: Red Hat
Component: RFEs
Version: 3.6.7
Hardware: All
OS: Linux
low
low
Target Milestone: ---
: ---
Assignee: Rob Young
QA Contact: Gil Klein
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2016-07-12 16:51 UTC by Robert Scheck
Modified: 2019-04-14 12:41 UTC (History)
7 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed:
oVirt Team: Docs
Target Upstream Version:


Attachments (Terms of Use)

Description Robert Scheck 2016-07-12 16:51:46 UTC
Description of problem:
Via bug #1063343 a functionality to report the total RX/TX byte statistics
for hypervisor and VM network interfaces was implemented (thanks!). The
official RHEV documentation however does not state if or when the counter
is overflowing. Technically, there must be an overflow somewhere, even it
is not documented so far.

Thanks to GSS we were able to track down that "bigint" data type (8 bytes)
is used in PostgreSQL database, thus the range from -9223372036854775808 to 
+9223372036854775807 is supported - that's 8 exabyte, if I did math right.

The question is what is going to happen if +9223372036854775807 is reached,
on the RHEV API level and on the database level. Will it reset to 0 (zero),
or will it switch to -9223372036854775808? What will the RHEV API report?

Version-Release number of selected component (if applicable):
RHEV 3.6.7

Actual results:
Technical max. value and possible value overflow behaviour of total RX/TX
byte statistics is not documented.

Expected results:
Documentation of the technical max. value and the possible value overflow
behaviour of total RX/TX byte statistics. Based on that, every customer can
at least calculate themself it it's realistic to reach this barrier or not.

Comment 3 Robert Scheck 2016-07-13 08:54:23 UTC
Yaniv, not sure what you expect by me (I am not a Red Hat employee, thus if
there was any internal comment with your needinfo flag, I did not receive it).

Comment 4 Yaniv Kaul 2016-07-13 09:01:44 UTC
(In reply to Robert Scheck from comment #3)
> Yaniv, not sure what you expect by me (I am not a Red Hat employee, thus if
> there was any internal comment with your needinfo flag, I did not receive
> it).

OK, did not know that - I suggest opening RFEs via GSS, which knows what the correct procedure is to open a RFE against RHEV.

My initial comment however was more towards 'who cares' - if your calculation or mine are correct, we'll overflow in ~100 years. It's worth ensuring the math is more or less correct (again, with GSS), but if it is, does it matter?

Comment 5 Robert Scheck 2016-07-13 09:24:18 UTC
Sorry, I maybe was initially unprecise. I more care about having the max.
value documented rather the exact overflow behaviour. What is also still
unclear is if the RHEV API's max. value is lower than 8 EB or not, given
we only talked about the database schema so far.

The exact things I would like to see are quite simple (and easily done):

a) Please enhance table 14.9 (Host NIC statistic types) to mention the max. 
   value in the column after data.total.rx, data.total.tx, errors.total.rx, 
   errors.total.tx.

b) Please enhance table 15.7 (Virtual machine NIC statistic types) to
   mention data.total.rx, data.total.tx, errors.total.rx, errors.total.tx
   at all (they are currently missing, while they technically exist) and 
   mention their max. values in the column after as well.

The RHEV API documentation, I am referring to, is:

  https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Virtualization/3.6/html-single/REST_API_Guide/index.html#sect-Network_Interfaces_Sub-Collection

Btw, while it's indeed unlikely to overflow soon per virtual machine, the 
overflow at the hosts should happen "way before" (no, I did no calculation 
here).

Cross-filed case 01256408 on the Red Hat customer portal.

Comment 8 Steve Goodman 2018-09-12 12:35:03 UTC
In conversation with Ryan Barry, setting the priority and severity to low.


Note You need to log in before you can comment on or make changes to this bug.