Bug 1392789 - Replace vmUUID to vmName in vm metrics' prefix
Summary: Replace vmUUID to vmName in vm metrics' prefix
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: vdsm
Classification: oVirt
Component: Core
Version: ---
Hardware: Unspecified
OS: Unspecified
medium
medium
Target Milestone: ovirt-4.1.0-alpha
: 4.19.2
Assignee: Yaniv Bronhaim
QA Contact: Lukas Svaty
URL:
Whiteboard:
Depends On:
Blocks: oVirt-Metrics-and-Logs
TreeView+ depends on / blocked
 
Reported: 2016-11-08 09:18 UTC by Yaniv Bronhaim
Modified: 2017-04-27 09:36 UTC (History)
11 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2017-04-27 09:36:12 UTC
oVirt Team: Infra
Embargoed:
rule-engine: ovirt-4.1+
bmcclain: priority_rfe_tracking+
ybronhei: devel_ack+
pstehlik: testing_ack+


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
oVirt gerrit 66234 0 None None None 2016-11-27 11:58:58 UTC

Description Yaniv Bronhaim 2016-11-08 09:18:50 UTC
Description of problem:
Currently metric's prefix includes vm uuid. This info is less important than the vm name which is known by libvirt as well. The uuid is internal to engine and less consumed by reports.
In addition, the metric length must be limited to 63 characters. Engine allows vmName up to 64 characters. As part of changing the uuid to name, trim the name to 30 bytes (the rest of the fields don't pass 33 chars in all cases).

Comment 1 Shirly Radco 2016-11-08 09:52:00 UTC
Since naming conventions may change from one env to another.
It might be better to trim the middle of the vm name, leaving the start and end characters with '...' between.

Comment 2 Sandro Bonazzola 2016-12-12 13:54:58 UTC
The fix for this issue should be included in oVirt 4.1.0 beta 1 released on December 1st. If not included please move back to modified.

Comment 3 Lukas Svaty 2017-01-30 14:32:33 UTC
Hi, can you provide verification steps here please?

Thanks,
LS

Comment 4 Yaniv Bronhaim 2017-02-13 12:41:54 UTC
you should be able to see vdsm sending metrics to port 8125. its enabled by default (eg. should be enough to run tcpdump -i lo 'port 8125' -v -X). there vdsm used to send metrics every 15seconds like vms.[UUID].mem... after this patch you should be able to see vms.[vm-name].mem... if the host runs vms.

Comment 5 Lukas Svaty 2017-03-30 12:12:52 UTC
verified in latest metrics vdsm-4.19.7-1.el7ev.x86_64

14:11:20.709025 IP (tos 0x0, ttl 64, id 28440, offset 0, flags [DF], proto UDP (17), length 71)
    localhost.41308 > localhost.8125: [bad udp cksum 0xfe46 -> 0x3d86!] UDP, length 43
	0x0000:  4500 0047 6f18 4000 4011 cd8b 7f00 0001  E..Go.@.@.......
	0x0010:  7f00 0001 a15c 1fbd 0033 fe46 766d 732e  .....\...3.Fvms.
	0x0020:  7465 7374 5f76 6d5f 6c73 7661 7479 2e62  test_vm_lsvaty.b
	0x0030:  616c 6c6f 6f6e 2e74 6172 6765 743a 3230  alloon.target:20
	0x0040:  3937 3135 327c 67                        97152|g
14:11:20.709073 IP (tos 0x0, ttl 64, id 28441, offset 0, flags [DF], proto UDP (17), length 70)
    localhost.41308 > localhost.8125: [bad udp cksum 0xfe45 -> 0xfa37!] UDP, length 42
	0x0000:  4500 0046 6f19 4000 4011 cd8b 7f00 0001  E..Fo.@.@.......
	0x0010:  7f00 0001 a15c 1fbd 0032 fe45 766d 732e  .....\...2.Evms.
	0x0020:  7465 7374 5f76 6d5f 6c73 7661 7479 2e6e  test_vm_lsvaty.n
	0x0030:  6963 2e76 6e65 7430 2e72 785f 6572 726f  ic.vnet0.rx_erro
	0x0040:  7273 3a30 7c67                           rs:0|g
14:11:20.709120 IP (tos 0x0, ttl 64, id 28442, offset 0, flags [DF], proto UDP (17), length 68)
    localhost.41308 > localhost.8125: [bad udp cksum 0xfe43 -> 0xe81f!] UDP, length 40
	0x0000:  4500 0044 6f1a 4000 4011 cd8c 7f00 0001  E..Do.@.@.......
	0x0010:  7f00 0001 a15c 1fbd 0030 fe43 766d 732e  .....\...0.Cvms.
	0x0020:  7465 7374 5f76 6d5f 6c73 7661 7479 2e62  test_vm_lsvaty.b
	0x0030:  616c 6c6f 6f6e 2e63 7572 3a32 3039 3731  alloon.cur:20971
	0x0040:  3532 7c67                                52|g


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