Mt. Gox reopens in 1 hour. Predictions?

[superduh]

mtgox.. is reopening slowly.
guess what.!!
still plenty of .01 btc trades going through. what is wrong with gox

[superduh]

disable ALL APIS please for the next 24 hours
but does gox ever listen

[ShireSilver]

Did they have the trading front end running on the same box as the dbase?  (As if they shouldn't have put that into 2 boxes already).

Maybe they went out and bot those PCIe Ram Disk cards?  (With 500K/day of revenue, shouldn't they have done that already).

The fact that they CAN upgrade the dbase in 12 hours scares me to begin with!

IT guys thoughts?

I thought I saw on another thread that yes, they did have the front end and the DB on the same box. Moving the DB to another machine *should* be a trivial task, but when dealing with mission critical components making sure it all works properly means caution and testing.

If the database upgrade is a simple swap for more power (better CPUs, more RAM, faster drives) then yes it can be done in 12 hours.

[12 years as a programmer, with the last 5 of them working on a large web platform.]

[tickets]

well, it's trading between 55 and 80 right now on various other exchanges....  so there would be an arb opportunity to sell if the gox bids were significantly higher than say 70, and buy back on other exchanges for short profit.    The arb-bots and arb-traders should have some fun here.

MtGox's posted reason for halting trade was  "to allow the market to cooldown following the drop in price"... they later said they would be upgrading their db servers.

To me, this says they were trying to support the price much as trading curbs kick in on NYSE during a crash.

There could be some surprises on the order book also. Remember things are very different now than after the first bitcoin bubble pop, with VC and other investors indicating interest.  If you're a VC and ready to buy bitcoin at 200, are you going to lose interest in buying it below 100?  

The Vinklevii twins announced (around 2pm?) they are major backers of bitcoin, and said they're buying the  dip...

I wouldn't doubt that during the halt, some phone calls were made, and buyers solicited to prop up the price.

I wouldn't be surprised to see big bids supporting price at major levels, say $60...  

Finally, there should be heavy load and wild swings for a while. At some point it will be clear whether the new db can handle the load, after which the lag will diminish, and prices will stabilize after a classic consolidation triangle pattern.

Which brings us to tomorrow, and next week.

There are 15k users backlog in the mtgox verification queue.  Most of these represent small time speculators or investors who were eager to buy at $200.   Some of these will chicken out... but they don't already hold coins.  Others will see the dip as an opportunity and go ahead with their buying.   But overall, there will still be people standing in line to buy, which should begin price rising again....

How's that for a guess?

[Malawi]

Everything will go smoothly.

Thats all the confirm i need.  Mt Gox will trade for 10 minutes and then lag, then sell off, then suspend trading.

What kind of dbase upgrade could they have done in 12 hours.  Swapped out the mechanical to ssd in their raid array? (As if they should have done this already)

Thoughts?  Was the dbase access the bottleneck?  Did they have the trading front end running on the same box as the dbase?  (As if they shouldn't have put that into 2 boxes already).

Maybe they went out and bot those PCIe Ram Disk cards?  (With 500K/day of revenue, shouldn't they have done that already).

The fact that they CAN upgrade the dbase in 12 hours scares me to begin with!

IT guys thoughts?

They are probably rushing in more hardware and possibly bandwith.

It's really something you down wanna do while systems are running at full capacity, as it can easily introduce errors.

If they actually have used only one machine, the trading software is likely to be designed for that. If so - my bet would be that they have brought in a new machine that can handle a lot more traffic.

Best case scenario would be that they have a system that is scalable, and that they just needed some time for installment/testing.
Guess they can afford a blade-stack by now.:-)

[Wuji]

Opening around $80 and then down under $50 tomorrow with more wild fluctuations throughout the day.

[doobadoo]

Did they have the trading front end running on the same box as the dbase?  (As if they shouldn't have put that into 2 boxes already).

Maybe they went out and bot those PCIe Ram Disk cards?  (With 500K/day of revenue, shouldn't they have done that already).

The fact that they CAN upgrade the dbase in 12 hours scares me to begin with!

IT guys thoughts?

I thought I saw on another thread that yes, they did have the front end and the DB on the same box. Moving the DB to another machine *should* be a trivial task, but when dealing with mission critical components making sure it all works properly means caution and testing.

If the database upgrade is a simple swap for more power (better CPUs, more RAM, faster drives) then yes it can be done in 12 hours.

[12 years as a programmer, with the last 5 of them working on a large web platform.]

Shouldn't it be as trivial as installing the same dbase server software package on the new machine, popping 10 gigbit enet cards in there, pluggin it into your router and then changing one line of code that contains the address of the server?

I did a complex street algo in SAS and MSSQL that processed millions of records, you never knew when the half dozen or so dbases i'd submit queries/updates to would move to a new address.   So you write the fucker "high Level" .  I queried a local table for the address of each dbase i would hit, as even some times the dbases themselves moved to different servers while some stayed.

Maybe i oversimplify, thats probably not how it queries.

[smoothie]

PREDICTION: LAG LAG FAG LAG

[yocko06]

I hope it hits $1. Then I will buy all the bitcoins I can

[superduh]

large purchases coming in at 129.. what a good day for people to arbitrage if they could

[doobadoo]

And it was really fun when the dumbass systems guys set up the new server and when migrating the dbase changed the schema.

[crazy_rabbit]

the forum lags the heck out so we can't keep each other company.

[Dr.Steve]

price will rise
price will fall

rinse and repeat

[iCEBREAKER]

chaos is my only prediction.

^ ^ ^

[justusranvier]

One would hope they had improvements ready to go for a while, but were just waiting for a chance to halt trading in order to make changes which could not easily be done while operating, like moving the trading engine to a separate machine from the web server and reconfiguring various monolithic components to be setup for failover/replication.

That's just hope though. My current expectation is that it will be a clusterfuck but I'd like to be pleasantly surprised.

[thoughtfan]

It will probably crash well below $100, probably in the range of $20-50.

This is how the mainstream press is viewing this event:
http://www.naturalnews.com/039880_bitcoin_bubble_panic_selling_accounts_frozen.html

Quote from the article:
"Bitcoin is now officially a failed experiment."

naturalnews is mainstream media? humph... news to me.

I was thinking that.  Some mainstream might have the same overall message but that reads like amateur journalism.  The 'officially a failed experiment' bit illustrates more than anything the writer's thinking if a statement is made emphatically enough it makes it true 

[crazy_rabbit]

Whomever has been running the DDOs in the past has their botnet on a clockwatch.

[proudhon]

price will rise
price will fall

[nandika]

my prediction is that they will get soo much traffic, so many refreshes, they won't be able to handle it...

[doobadoo]

...and reconfiguring various monolithic components to be setup for failover/replication.

Failover?!?!  Redundancy!?!?  Dude this is GOX we're talking about.  The company that sits there and opens up thousands of accounts without knowing whether their platform can even handle it!