Three releases this weekend

[Steve Sokolowski]

There will be three releases this weekend.

In the first release, which will occur today, the default payout coin will be changed to litecoin. New customers will be paid in 100% litecoins, and existing customers will be paid in litecoins if all of their preferred coins are in error. The payout threshold for litecoins is $1, so it will be easy to pay customers who accumulate litecoins due to an error, unlike the current situation where customers earn $1 in useless bitcoins when these issues occur. This release will not require downtime and the only way anyone will notice it is if a bug is discovered for the first time in production and a revert is necessary.

In the second release, we will be changing the mining server to use pypy instead of CPython, as well as releasing other performance improvements. The mining server will then be limited to a total of 1.6TH/s estimated, up from the current estimated 1.05TH/s. This almost meets the requirement of about 2TH/s I wanted to get to before SHA-256. This release was scheduled for July 2 last week and will continue to be on that schedule. This one requires up to 30 minutes of downtime because it involves installation of many packages and the new python compiler. It includes WAMP optimizations to try to make the system respond more quickly to WAMP function calls.

In the final release, we will release performance improvements to the webserver. The website is often pegged at 100% CPU and 8GB memory, and we haven't ever optimized it, because it was never a bottleneck. This weekend will mark the start of optimizations in that area. Since releases of the website are low-risk and easily revertable, Chris will do it alone on Monday.

[FRISKIE]

New customers will be paid in 100% litecoins

Does it mean that new customers will not be able to select any other coins for payout besides Litecoin?

[fattygrub]

Steve,

On the second release, can you guesstimate what date/time the 30 minutes downtime on the mining server will be?

[FRISKIE]

And a question regarding the 2nd release:

This sounds like a major release and 30 minutes downtime can easily turn into longer, with likely disruptions and unintended effects to work out.

The system is currently running the best it has ever run in the months I have been here.

Considering what we have all been through with the DDoS downtime and lost revenue, don't you want to consider leaving the system AS IS for a while and let us all earn a little profits?

And just as important - Regain some confidence?

[FRISKIE]

Or let me put it this way.

I don't believe that you guys have understood very well that you are providing a SERVICE to your customers, and as such you need to put the concepts of stable service delivery ahead of all else!

This means Change Management which considers risk of service impact, and of course the projected downtime if everything goes to plan.


In any sane organization I have ever worked in before, Service Delivery Management would NEVER approve a MAJOR Change release so soon after a Major Incident had caused 2 1/2 days of downtime!

Especially if the systems were all functioning well and delivering a high value service again (finally)

The question to you guys now is - "Do you want to begin run the site according to best industry practices?"

If yes, you need to slow it down a bit and begin to view the site as a business, and consider the Service Delivery aspects first and foremost.

[Eyedol-X]

To me this looks like a situation where they are not letting outages beyond their control (Ddos) affect a planned roadmap for rollout. To me it's not a big deal because while the outage could mean an additional outage, it's not like our miners do not have a fail over feature that can be used to mine on a different pool.

I could understand if the outage was caused by their error but that doesn't appear to be the case here with the information that's available.

I don't see the need to delay previously planned outages just because there was an outage a couple of days ago that had nothing to do with their servers not performing properly. It could have just as easily been an ISP outage causing 2 days of downtime, I don't think that should cause a release delay.

[FRISKIE]

@ Eyedol-X - My point was that in larger successful organizations this would be taken into account from a "customer and service delivery" perspective.

You have valid points as well, it's simply a matter of which perspective the brothers prefer to look at it from.

[fattygrub]

RE: Scheduled mining server downtime

At least a countdown/warning would be handy...e.g.

T-4hrs..T-30min

[AvPro]

I would concur with FRISKIE. It seems after best part of a week we're at a stable zone , especially yesterday. (With the exception of BCC playing up).

Would appreciate to roll with the current stability for a while longer before any new possible points of downtime are inadvertently introduced.

My 2 cents

Thanks