cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
9202
Views
10
Helpful
9
Replies

SYS-3-CPUHOG = SAUtilReport

Hi peeps,

 

We're seeing this CPUHOG after a system reboot and wanted to ask if anyone know if I should contact TAC or if there are any other solutions? Anyone knows what this SAUtilReport does?

 

System:
Model: C9200-24P

FW version: 17.03.03

 

Reboot reason might be interesting:

System returned to ROM by Critical process linux_iosd_image fault on rp_0_0 (rc=139) at 12:22:18 UTC Mon Aug 23 2021
System image file is "flash:cat9k_lite_iosxe.17.03.03.SPA.bin"
Last reload reason: Critical process linux_iosd_image fault on rp_0_0 (rc=139)

 

There is a crash log, but it's quite big, so not sure if it makes sense to upload here.

 

BR

Claus

1 Accepted Solution

Accepted Solutions

Fixed it by following this workaround, in CSCvv72609 :
#license smart factory reset
#reload

Other processes which were affected:
SAGetRUMIds
SAUtilRepSave

View solution in original post

9 Replies 9

balaji.bandi
Hall of Fame
Hall of Fame

if this is imacting, raise an TAC case, if not try difference version like 17.3.4 or above ?

 

BB

***** Rate All Helpful Responses *****

How to Ask The Cisco Community for Help

Leo Laohoo
Hall of Fame
Hall of Fame

@Claus Juhl Pedersen wrote:

Last reload reason: Critical process linux_iosd_image fault on rp_0_0 (rc=139)


Raise a TAC Case and get them to confirm if this is CSCvq78529. 

WARNING:  Make sure to raise a TAC Case during the hours of operation of Europe or RTP/South America time.  

you know how much it costs to raise a tac case? is it worth it just to let them confirm it may be hits some bug and get their valuable advise to upgrade the ios? 


@tommar wrote:

you know how much it costs to raise a tac case? is it worth it just to let them confirm it may be hits some bug and get their valuable advise to upgrade the ios? 


IOS-XE, aka Polaris, is buggy and leaks like a sieve.  

I have seen and heard of old platforms such as 6000/6500 in the core with uptimes of >5 years without any issues.  But for Polaris, the only way, that I can think of (and based on experience), to minimize the chances of crashes is to schedule regular reboot.  For 3650/3850, it would be once every 3 months.  If the switches are on 16.12.X, then every four weeks.  For the Catalyst 9k switches, once every 9 to 12 months.  

I have been exposed to Cisco OS since 2005 and I have never had so many TAC cases created since the introduction of IOS-XE.  It is laughably easy to find bugs with no experience necessary.  

And a lot of the bugs we have found are "0-config":  Take a switch, router or 9800 out of the box, load the firmware and *bang*, bug is triggered.  

Fixed it by following this workaround, in CSCvv72609 :
#license smart factory reset
#reload

Other processes which were affected:
SAGetRUMIds
SAUtilRepSave

jmcgrady1
Level 1
Level 1

Cisco's software is getting steadily worse. Reading CSCvv72609; the workaround, which will clear the backlog, requires that the smart licensing be wiped - which could trigger a crash! And even if it works, the switch must be restarted,

Honestly, they don't have a graceful way of restarting an errant process?


@jmcgrady1 wrote:

Cisco's software is getting steadily worse.


It has been and it will only get worst moving on.  

There is really no need to get a switch, with multi-CPU, SD-WAN or SD-Access, if it is will only be doing simple Layer 2 packet shuffling.  

Leo what is a good choice for a switch?  My needs are layer 2 switching (1000mb/s), 4 SFP (10gb capable is a bonus), PoE, trunking. Stacking is a bonus - if the stack can reboot each member 1 at a time for things like firmware update. For some switches, i need layer 3 capability: EIGRP, QoS.


@jmcgrady1 wrote:

My needs are layer 2 switching (1000mb/s), 4 SFP (10gb capable is a bonus), PoE, trunking. Stacking is a bonus - if the stack can reboot each member 1 at a time for things like firmware update. 


Layer 2, 802.1q trunking, stacking ... Nothing wrong with Catalyst 1000.  

At the very least Catalyst 1000 does not support Cisco "Smart" (uh-huh, suuuuuure) License.

QoS in a network is redundant.  QoS on the WAN-facing port is a wasted effort.