rssLink RSS for all categories
 
icon_red
icon_green
icon_red
icon_red
icon_blue
icon_green
icon_green
icon_red
icon_red
icon_red
icon_orange
icon_green
icon_green
icon_green
icon_green
icon_blue
icon_green
icon_orange
icon_red
icon_green
icon_red
icon_red
icon_green
icon_red
icon_red
icon_red
icon_red
icon_orange
icon_green
 

FS#1864 — FS#5840 — pcc-22 et pcc-25

Attached to Project— Dedicated Cloud
Incident
Backend / Core
CLOSED
100%
We have just detected 2 switch down.
WE intervene.
Date:  Thursday, 29 September 2011, 16:48PM
Reason for closing:  Done
Comment by OVH - Thursday, 29 September 2011, 16:26PM

pcc-25 has come back.
pcc-25 was not completely planted but partially
so it has not cut the hosts ports.
We have cut the pcc-22 port.The traffic has been taken by the PCC-23 card on the other network hosts.


Comment by OVH - Thursday, 29 September 2011, 16:36PM

PCC-22 has a stable version of NX-OS but to get
around a bug that grubs the CPU of N5 we
have enabled the spantree.

Since we have received an alpha version of NX-OS which we are using in all the other pcc-XX-n5
and which allows us to protect the CPU to some
measure. waiting the beta next week.



Comment by OVH - Thursday, 29 September 2011, 16:47PM

we have found the origin of the problem which was due to a
human error and a bug: a technician took
an IP already in production pcc-25 for another
equipment which has caused the alert on the PCC-25 that has
not planted and does not degraded the service. the
technician also took the IP of the PCC-22 and plated the switch (!)

it has been updated with the version
N5000-uk9.5.1.3.N1.0.254.bin
we turned the spantree. we protect the CPU against
the BPDU packets that we have on the network and
are processed in software.




Comment by OVH - Thursday, 29 September 2011, 16:47PM

cc-22-n5(config-if-range)# sh proc cpu sort

PID Runtime(ms) Invoked uSecs 1Sec Process
----- ----------- -------- ----- ------ -----------
3206 211 139 1519 20.9% netstack

20%-30% for software treatment.