Wed, 27 Sep 2006
I've got a message from Dan saying
that Odysseus is strangely slow,
he said that copying files to Odysseus simply hanged after few kilobytes.
I have tried to run
tcpdump, and it seemed that Odysseus
did not send reply to some TCP frames. After figuring out that copying
data did not work even from my workstation to Odysseus, I started to suspect
the Cisco switch, in which I have upgraded the firmware yesterday.
Yesterday I have upgraded the firmware in some of our Cisco switches (in order to finally get SSH working on them). After looking at MRTG graphs for Odysseus, I have seen something like this (it was before 2pm today):
So for some reason, the link speed negotiation between Odysseus
and the Cisco switch has ended up at 100 Mbit/s FD, instead of 1 Gbit/s FD.
After restarting the negotiation using
ethtool it works OK.
I hereby declare Cisco 3750 being a total crap. For example, while other switches can be rebooted in a minute or so, the boot of 3750 takes much longer, and even then, it takes another half a minute for the ethernet interfaces to become active (and it is this way even for newly plugged-in cables). Stay away from Cisco switches (at least for L2 switching, I am not familiar with their L3 gear). HP is much more open, supported, and generally better.
Sorry for the long delay between my previous post and this one. I have been off-line for 10 days, and I have been busy catching up with my mail queue since then.
14 replies for this story:
davro wrote: insufficient knowledge causes wrong results
use switchport portfast in interface config. Cisco conform exactly to specifications. All ports must be blocked for 30 sec. in learning state when running spanning tree.
little mistake, use spanning-tree portfast.
BTW, HP lacks many features, which are very useful in large L2 networks (PVST+, VTP, error recovery, UDLD, ...)
And did you read the documentation? It's publicly available on vendor website. I know, it's a bit longer... but deprecate something without configuration knowledge shows your disability, not some problems with that box... And by the way - c3750 is Layer3 switch... :)
Yenya wrote: Re: for davro
Hmm, bad specification then, I would say. No other vendor's default configuration does this 30-second delay. But thanks for letting me know the workaround. Funny you mention standard compliance, and then a proprietary thing like PVST+ or VTP as a Cisco advantage. We have used VTP in the past, but it was pain in the a** - for example in some of our Catalyst switches it was configured separately, outside the main config (in a "vlan database" command, IIRC). So saving a config to the TFTP server actually did not save the full state of the switch. Ugly as hell.
Yenya wrote: Re: for Danny
The Cisco documentation is one of the most disorganized pieces of data I am aware of (and I have been using various Cisco products for at least 7 years now). 3750 can do L3 switching only with a special firmware image, which we do not have. I may repeat all the Cisco drawbacks here (like no free software upgrades, messy documentation, proprietary protocols, configuration hacks like the ugly "vlan@community" hack instead of Q-BRIDGE-MIB on 4912 and 2948G, etc.). But this was not a point of my blog post. The point was that it has failed to do a correct speed autonegotiation after the reboot, and that the reboot took an exceptionally long time.
30 sec. delay is failsafe default. If you don't do this, you can be suprised by spanning tree loops. Lovely to trace and debug. And if some vendor's product doesn't do this, then it fails to conform to specs (so it is a crap). And those mentioned PVST+, VTP are optional, you can easily use standard MSTP (a bit stupid IMHO) or rapid SPT.
BTW: 3750 is L3 switch even with standard image. It lacks support for OSPF and other advanced L3 features.
Yenya wrote: Re: for davro
As for 3750 being an L3 switch, I did not know that. Thanks. The 30sec delay: maybe the prefered solution would be to not run any spanning tree algorithm at all by default. Those who want it can enable it manually.
Yes, you can turn SPT off, but I don't think it is a way to go. It is not a safe solution. You can run into deep troubles. And 3750 is a switch for bigger networks, not a "single switch" solution, so SPT should be turned on by default.
30sec delay (by default turned STP/PVST) is standard thing on all of Cisco switches in the long term. I do not understand, why are you surprised by this thing if you say something like "I have been using various Cisco products for at least 7 years now"... this is nothing new and "specific" to Cat3750... I personally think much about your work in Linux community, but this your blogpost about Cisco gear is absolutely amateur.
hi yenya do u know way by which we can emulate the cisco switch hardware. i mean are there any hardware or firmware debugging tools that we can use to find the internal working of the hardware as to what hardware checks it makes before it loads the ios into it. regards sebastan
I hereby declare that you have to study hard first before saying that something is crap.
Yenya wrote: Re: nonbeginner
Well, I was wrong in one thing (the spanning-tree portfast problem), but long boot times, non-free firmware upgrades, and occasional autonegotiation failure remain as problems of Cisco. Feel free to enlighten me how the remaining problems do not mean Cisco 3750 is a crap.