jgherbert

Absolutely agreed. The funny thing is that in my (possibly biased) experience it is often the network team who pull together the silos that make up a service, and track down the issue. Maybe that's because Mean Time To Innocence (MTTI, as rschroeder​ mentioned above) seems to be a KPI for the network team, more than any…

I could put that on most of my posts, tbh

I wonder if Tholarwinds will see an opportunity here? ;-)

Part of the problem is the whole "boil the ocean" paradigm. And during those ridiculous amounts of work and planning and coordination, nothing is working and there's a danger that support for the initiative wanes. To your point about getting in at the planning layer, that's exactly it; if you build something out from the…

That's another argument put forward for why we won't automate ourselves out of a job; there's always something else you can do with the time you free up!

The internet is a wonderful thing for casual coders. My only word of caution is that if you don't know what the code does, it's going to be hard to troubleshoot and integrate into your own program. Modules/Packages are great for addition functions on somebody else's dime, so to speak, but code snippets need to come with a…

This kind of thing always makes me wonder how much we're missing...

*lol* I hope it's not a retread by me! At the least, I can promise it wasn't plagiarized: it was written without reference to any other materials on the subject. Glad to hear we're on the same track though!

Thanks, Vinay.

prowessa‌> I probably should have added in the post, "and turn off local logging / buffering of those logging streams or your device is probably gonna be toast ;-)"

I'm a control freak too, but avoiding DHCP in general is way too time consuming and doesn't really scale, IMO.

The vendors already win, in spades, when they sell a piece of hardware. Conversely though, if an industry standard emerges and vendors start losing RFPs because they don't support it, they'll push it back up the priority list (I guess that counts as enough people pushing for it). This is perhaps the third comment about…

*grin* True up to a point, so long as the language you know best also supports what you need to do. On the other hand, if you aren't currently familiar with a language and are starting from a clean slate, so to speak, what's the right one to choose? Obviously I have my opinions as stated above, but to echo your sentiment,…

You make a good point; I did not mention "and document them thoroughly with good examples that will help other people use them," which is a failing on my part ;-)

Well, ironically enough somebody might not be using all those nodes (the abandoned compute resources I mentioned at the end) ;-) We can't help being in the weeds most of the time; we're here because we are expert horticulturists. That said, being able to see all of the weeds and not just looking at our favorite species of…

As with my last post on SNMP, pervasive != best. In particular for firmware downloads I'd argue that TFTP is an absurd protocol, not performing a checksum on the data until the entire transfer has completed. My real point, perhaps not made clearly, was that for TFTP to be effective we still really need a way to get…

familyofcrowes‌ You've hit the nail squarely on the head there. Assuming you can generate all those logs, storing and processing them can require huge resources. Retention times are also frequently set for legal rather than engineering reasons, which can mean a SAN (or similar) might end up being a requirement for the many…

There may be something in what you say... :-)

Is that $19 per leak, or do you get unlimited leaks for that fee?

Baby steps perhaps? Have you ever played with Lego Mindstorms? The standard programming interface is exactly as you describe - dragging boxes around and letting the program build the code for you. On the other hand, it's not magic; if you can't figure out the logical steps and conditions in order to decide what action(s)…

Where's the key kept?

"its popularity stems from how well it works the majority of the time for most applications." --Flies swarm around cow dung. Conclusions may thus be drawn ;-) "SNMP has been in most cases a reliable and solid protocol" --I'm going to argue that SNMP is the lazy option. It works within reason, but it doesn't scale…

At least until the pictures were tweeted out, right cahunt‌?

If the management software supports it, any migration of protocols should be as simple as checking the SNMP box instead of the SomeNewProtocol, right? So it wouldn't be something you'd have to hate to change, as such, hopefully it would be a smooth migration, and for a period of time you would have a mixture of old and new…

"Silo Of Empires: Turf War. Play online now." Offers in-career purchases.] Yeah, that'll do it.

Very true. It may be a case of figuring out how to represent that kind of indirect dependency - or whether to simply list them as if they were direct, e.g. ApplicationA is dependent on MQ and a dependency on ApplicationB, since failure of either of those elements will cause problems for ApplicationA. It's certainly…

Thanks, Steve. I like the analogy of SNMP as STP; I shall try to remember it. "Ubiquitous, but suboptimal." ;-) There is so much going on in the typical network now, each requiring its only special kind of management system, it's really hard to look over all those systems and make sense of the various data sufficient to…

I like that too, but... well, see my next post. All smells like screen scraping to me.

Yay, job security!