General list for discussing Bufferbloat
 help / color / mirror / Atom feed
* [Bloat] RPM for sqm-scripts
@ 2019-07-31 19:32 Kenneth Porter
  2019-07-31 19:37 ` Sebastian Moeller
  2019-08-08 14:07 ` Toke Høiland-Jørgensen
  0 siblings, 2 replies; 4+ messages in thread
From: Kenneth Porter @ 2019-07-31 19:32 UTC (permalink / raw)
  To: bloat

I decided to try the sqm-scripts and created an RPM to make it easy to 
install/upgrade/uninstall on my CentOS 7 system. Here's the result:

<https://sewingwitch.com/ken/Stuff/sqm-scripts/>

An initial attempt at using simple.qos on my system killed performance and 
the interface became unusuable, but I was probably too optimistic in 
setting the bandwidth numbers based on best response from speedtest-cli. I 
temporarily disabled it until I can test during off-hours.


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [Bloat] RPM for sqm-scripts
  2019-07-31 19:32 [Bloat] RPM for sqm-scripts Kenneth Porter
@ 2019-07-31 19:37 ` Sebastian Moeller
  2019-07-31 20:01   ` Kenneth Porter
  2019-08-08 14:07 ` Toke Høiland-Jørgensen
  1 sibling, 1 reply; 4+ messages in thread
From: Sebastian Moeller @ 2019-07-31 19:37 UTC (permalink / raw)
  To: Kenneth Porter; +Cc: bloat



> On Jul 31, 2019, at 21:32, Kenneth Porter <shiva@sewingwitch.com> wrote:
> 
> I decided to try the sqm-scripts and created an RPM to make it easy to install/upgrade/uninstall on my CentOS 7 system. Here's the result:
> 
> <https://sewingwitch.com/ken/Stuff/sqm-scripts/>

	Pretty cool! (Caveat: due to lack of a centos system I will not be able to actually test it)


> 
> An initial attempt at using simple.qos on my system killed performance and the interface became unusuable, but I was probably too optimistic in setting the bandwidth numbers based on best response from speedtest-cli. I temporarily disabled it until I can test during off-hours.

	So maybe run the following to get some debug output:

SQM_DEBUG=1 SQM_VERBOSITY_MAX=10 /etc/init.d/sqm start ; tc -s qdisc ; tc -d qdisc ; SQM_DEBUG=1 SQM_VERBOSITY_MAX=10 /etc/init.d/sqm stop

I would hope that this should undo its damage at the end but will also capture the wedged state in between. Then again this still might lead to an unusable interface....


Best Regards
	Sebastian

> 
> _______________________________________________
> Bloat mailing list
> Bloat@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/bloat


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [Bloat] RPM for sqm-scripts
  2019-07-31 19:37 ` Sebastian Moeller
@ 2019-07-31 20:01   ` Kenneth Porter
  0 siblings, 0 replies; 4+ messages in thread
From: Kenneth Porter @ 2019-07-31 20:01 UTC (permalink / raw)
  To: bloat

--On Wednesday, July 31, 2019 10:37 PM +0200 Sebastian Moeller 
<moeller0@gmx.de> wrote:

> Pretty cool! (Caveat: due to lack of a centos system I will not be able
> to actually test it)

It should work on any recent Red Hat system, including RHEL, Fedora, 
CentOS, and Scientific. It's using the systemd files so it can't start 
automatically with the older initscript-based versions. RH doesn't have the 
ifup hook for packages so it doesn't include that feature to enable/disable 
the SQM on interface up/down.

Note that this is just a packaging wrapper. I didn't change any of the 
scripts. This just drops all the files into the correct place and updates 
the package database so they can be easily removed or upgraded as 
necessary.

To install, just use "yum install sqm-scripts-1.3.0-1kp.noarch.rpm".

Start with "systemctl start sqm@eth0" (substituting your interface name for 
eth0). You need to first create /etc/sqm/eth0.iface.conf based on the 
provided example.

> 	So maybe run the following to get some debug output:
>
> SQM_DEBUG=1 SQM_VERBOSITY_MAX=10 /etc/init.d/sqm start ; tc -s qdisc ; tc
> -d qdisc ; SQM_DEBUG=1 SQM_VERBOSITY_MAX=10 /etc/init.d/sqm stop
>
> I would hope that this should undo its damage at the end but will also
> capture the wedged state in between. Then again this still might lead to
> an unusable interface....

Thanks, I'll give that a try, probably tomorrow morning before everyone 
gets in. I was able to quickly get in to repair the damage and stop the 
service ("systemctl stop sqm@em2") by remoting in through a server I have 
in parallel and shelling to the router's internal interface (which isn't 
being shaped). The slowness didn't happen instantly but after a minute or 
more, when I got a call from the office that "the internet is down!". The 
external interface wasn't completely dead, just extremely slow, enough to 
eventually kill my ssh session. It recovered instantly with the stop 
command.

Ken



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [Bloat] RPM for sqm-scripts
  2019-07-31 19:32 [Bloat] RPM for sqm-scripts Kenneth Porter
  2019-07-31 19:37 ` Sebastian Moeller
@ 2019-08-08 14:07 ` Toke Høiland-Jørgensen
  1 sibling, 0 replies; 4+ messages in thread
From: Toke Høiland-Jørgensen @ 2019-08-08 14:07 UTC (permalink / raw)
  To: Kenneth Porter, bloat

Kenneth Porter <shiva@sewingwitch.com> writes:

> I decided to try the sqm-scripts and created an RPM to make it easy to 
> install/upgrade/uninstall on my CentOS 7 system. Here's the result:
>
> <https://sewingwitch.com/ken/Stuff/sqm-scripts/>

Cool! Care to fix it up to publishable state (you have a TODO in there),
then we can try to get it into upstream Fedora? :)

-Toke

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2019-08-08 14:07 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-07-31 19:32 [Bloat] RPM for sqm-scripts Kenneth Porter
2019-07-31 19:37 ` Sebastian Moeller
2019-07-31 20:01   ` Kenneth Porter
2019-08-08 14:07 ` Toke Høiland-Jørgensen

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox