mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   mersenne.ca (https://www.mersenneforum.org/forumdisplay.php?f=173)
-   -   mersenne.ca (https://www.mersenneforum.org/showthread.php?t=23051)

storm5510 2018-02-12 17:35

mersenne.ca
 
I have not been able to contact James' site since yesterday afternoon. Does anyone have any ideas as to what the situation is?

Note: I looked through the posts and didn't find anything relative.

VictordeHolland 2018-02-12 17:44

Looks up from here?

retina 2018-02-12 17:49

[url]http://downforeveryoneorjustme.com/www.mersenne.ca[/url]

[quote]It's not just you! [url]www.mersenne.ca[/url] looks down from here.[/quote]

The redirect to [url]www.mersenne.ca[/url] is working, but from there it dies.

VictordeHolland 2018-02-12 18:53

1 Attachment(s)
Can access it on my phone a couple minutes ago

CRGreathouse 2018-02-12 19:04

It's down for me. Error message on Chrome:

[quote]This page isn’t working
[url]www.mersenne.ca[/url] unexpectedly closed the connection.
ERR_INCOMPLETE_CHUNKED_ENCODING[/quote]

chalsall 2018-02-12 19:06

[QUOTE=VictordeHolland;479889]Can access it on my phone a couple minutes ago[/QUOTE]

There seems to be something weird going on... Everything [URL="https://www.mersenne.ca/status/tf/0/7/3/8000"]appears to work[/URL] except for the top-level page.

Dubslow 2018-02-12 23:51

[QUOTE=chalsall;479891]There seems to be something weird going on... Everything [URL="https://www.mersenne.ca/status/tf/0/7/3/8000"]appears to work[/URL] except for the top-level page.[/QUOTE]

Nice catch, that's the behavior I see.

(Also this is totally the wrong forum.)

storm5510 2018-02-13 00:07

[QUOTE=Dubslow;479908]Nice catch, that's the behavior I see.

([B]Also this is totally the wrong forum.[/B])[/QUOTE]

You are correct. It is. Sometimes my brain is somewhat more mushy than others.

chris2be8 2018-02-13 16:38

Trying wget I found: [code]
chris@linux-5hwg:~/bin> wget 'www.mersenne.ca'
--17:43:59-- http://www.mersenne.ca/
=> `index.html'
Resolving www.mersenne.ca... 74.208.111.38
Connecting to www.mersenne.ca|74.208.111.38|:80... connected.
HTTP request sent, awaiting response... 418 You are a teapot
17:45:05 ERROR 418: You are a teapot.
[/code]And for https: [code]
chris@linux-5hwg:~/bin> wget 'https://www.mersenne.ca/'
--17:41:31-- https://www.mersenne.ca/
=> `index.html'
Resolving www.mersenne.ca... 74.208.111.38
Connecting to www.mersenne.ca|74.208.111.38|:443... connected.
ERROR: Certificate verification error for www.mersenne.ca: self signed certificate
ERROR: certificate common name `Parallels Panel' doesn't match requested host name `www.mersenne.ca'.
To connect to www.mersenne.ca insecurely, use `--no-check-certificate'.
Unable to establish SSL connection.

chris@linux-5hwg:~/bin> wget --no-check-certificate 'https://www.mersenne.ca/'
--17:41:55-- https://www.mersenne.ca/
=> `index.html'
Resolving www.mersenne.ca... 74.208.111.38
Connecting to www.mersenne.ca|74.208.111.38|:443... connected.
WARNING: Certificate verification error for www.mersenne.ca: self signed certificate
WARNING: certificate common name `Parallels Panel' doesn't match requested host name `www.mersenne.ca'.
HTTP request sent, awaiting response... 418 You are a teapot
17:43:01 ERROR 418: You are a teapot.
[/code]That error code is obviously a joke, but not a very helpful one.

Chris

heliosh 2018-02-13 17:24

[QUOTE=chris2be8;479960] [code]
17:43:01 ERROR 418: You are a teapot.
[/code]That error code is obviously a joke, but not a very helpful one.
[/QUOTE]
An official one though.
[URL]https://tools.ietf.org/html/rfc2324[/URL]
[CODE]Any attempt to brew coffee with a teapot should result in the error
code "418 I'm a teapot". The resulting entity body MAY be short and
stout.[/CODE]

Uncwilly 2018-02-13 23:32

Still broked. :(

storm5510 2018-02-14 13:44

James must have had a severe failure at some level. I would hate to think he has lost his entire database! :no:

science_man_88 2018-02-14 14:07

[QUOTE=storm5510;480030]James must have had a severe failure at some level. I would hate to think he has lost his entire database! :no:[/QUOTE]

[url]http://www.mersenne.ca/exponent/29[/url] works for me. Its probably just the main page.

chalsall 2018-02-14 14:26

[QUOTE=storm5510;480030]I would hate to think he has lost his entire database! :no:[/QUOTE]

He hasn't. As mentioned above, all the queries continue to work (for [URL="https://www.mersenne.ca/status/tf/0/30/3/8000"]example[/URL]), and were updated last night as expected.

Why his top-level-page is reporting the "Teapot" error is a bit strange, admittedly....

ET_ 2018-02-14 14:54

It works for me.

kladner 2018-02-14 22:44

[QUOTE=heliosh;479966]An official one though.
[URL]https://tools.ietf.org/html/rfc2324[/URL]
[CODE]Any attempt to brew coffee with a teapot should result in the error
code "418 I'm a teapot". The resulting entity body MAY be short and
stout.[/CODE][/QUOTE]
Big laugh at the link to the coffee pot protocol. :lol: [SIZE=4][B]HTCPCP!?!?! :missingteeth:
[/B][/SIZE]

chalsall 2018-02-14 22:53

[QUOTE=kladner;480072]Big laugh at the link to the coffee pot protocol.[/QUOTE]

If you enjoyed that, check out the [URL="https://tools.ietf.org/html/rfc2549"]IP over Avian Carriers with Quality of Service[/URL] RFC.... :smile:

Uncwilly 2018-02-14 23:12

[QUOTE=chalsall;480073]If you enjoyed that, check out the [URL="https://tools.ietf.org/html/rfc2549"]IP over Avian Carriers with Quality of Service[/URL] RFC.... :smile:[/QUOTE]
[URL="https://what-if.xkcd.com/31/"]Never underestimate the bandwidth of a station wagon full of tapes hurtling down the highway.[/URL]

Mark Rose 2018-02-14 23:23

[QUOTE=kladner;480072]Big laugh at the link to the coffee pot protocol. :lol: [SIZE=4][B]HTCPCP!?!?! :missingteeth:
[/B][/SIZE][/QUOTE]

I wrote a heavily used* API with that as an Easter egg. If you try to use the HTTP method BREW on the hidden /teapot endpoint it will error at you.

*only a few billion calls per month, nothing extreme.

kladner 2018-02-15 01:46

[QUOTE=chalsall;480073]If you enjoyed that, check out the [URL="https://tools.ietf.org/html/rfc2549"]IP over Avian Carriers with Quality of Service[/URL] RFC.... :smile:[/QUOTE]
Holy Birdpoop, Batman!

LaurV 2018-02-15 01:59

[QUOTE=kladner;480072]Big laugh at the link to the coffee pot protocol. :lol: [SIZE=4][B]HTCPCP!?!?! :missingteeth:
[/B][/SIZE][/QUOTE]
Yuck, no "Romanian" in the language list? :razz:
Hehe...

0PolarBearsHere 2018-02-15 12:29

[QUOTE=ET_;480040]It works for me.[/QUOTE]

Working here too. I'm assuming it got fixed.

ramgeis 2018-04-13 11:12

mersenne.ca seems to be down :(
Is there anyone who can still access it?

Chuck 2018-04-13 12:16

I can't access the site either. It went down Thursday night around 10:30 EDT and has not responded as of Friday morning.

storm5510 2018-04-13 13:00

[QUOTE=Chuck;485229]I can't access the site either. It went down Thursday night around 10:30 EDT and has not responded as of Friday morning.[/QUOTE]

Still down as of 9 AM EDST He has a vast amount of data stored, somewhere. I wonder what he runs all of this on?

Uncwilly 2018-04-13 13:37

Hopefully there is a backup or 3 in reliable hands.

petrw1 2018-04-13 15:09

[QUOTE=storm5510;485231]Still down as of 9 AM EDST He has a vast amount of data stored, somewhere. I wonder what he runs all of this on?[/QUOTE]

Can't be a Commodore 64 because mine never crashed.

Chuck 2018-04-13 19:35

Mersenne.ca is back online
 
Back up as of about 15:00 EDT :smile:

Chuck 2018-04-13 20:16

It was up for an hour but down again
 
It was up for an hour and I reported some results, but it is back down again.

James Heinrich 2018-04-13 20:24

Sorry guys, my server isn't feeling well. It keeps crashing for reasons not entirely clear. The data center just recently replaced one HDD but I hadn't yet had chance to even start rebuilding the array before the server went down again. I'm still trying to figure out where the problem lies and get it resolved.

James Heinrich 2018-04-13 23:30

Server may be down for a little longer than expected. There's something more wrong than an aging HDD, perhaps faulty RAM, not sure yet. Data center ticket opened, hope they can find and fix the problem soon.

[QUOTE=storm5510;485231]He has a vast amount of data stored, somewhere. I wonder what he runs all of this on?[/QUOTE]Depends on your definition of "vast", but yeah, it's about 600 million records taking up ~50GB (with indices). Of that, however, most of the data is replicating stuff already known from PrimeNet. The "important" bit is the TF effort above 1000M, which is not duplicated on any other server (but is backed up monthly to my home systems). Once the server is back online, if there's interest I can quite easily make this data available to anyone who might be interested (at a public URL, there's no secret data here).

Uncwilly 2018-04-13 23:42

Maybe Mike can keep a mirror of the data on the MF server. You also have a bunch of really useful tools, back them up too.

Chuck 2018-04-13 23:49

Friday 13th
 
I'd blame it on Friday the 13th, except for the fact that the problem started late Thursday night.

Uncwilly 2018-04-14 00:12

[QUOTE=Chuck;485269]I'd blame it on Friday the 13th, except for the fact that the problem started late Thursday night.[/QUOTE]Thursday where? Wasn't it Friday in Thailand already?

James Heinrich 2018-04-14 00:17

Where the server lives it was Thursday. Where you live is irrelevant :smile:

James Heinrich 2018-04-14 14:25

mersenne.ca is back online, at least for right now. Turns out the server was overheating (not my fault, I'm not running mprime on it or anything) and causing the crashing. They've swapped my drives over to a new physical server and it's back online. I still need to rebuild the RAID, so there might be a reboot or two coming today, but otherwise it should be back online.

It wouldn't hurt to re-submit any TF>1000M results from the last 3 days. They've probably been received, but just in case it doesn't hurt to resubmit them.

storm5510 2018-04-14 16:44

[QUOTE=James Heinrich;485267]Server may be down for a little longer than expected. There's something more wrong than an aging HDD, perhaps faulty RAM, not sure yet. Data center ticket opened, hope they can find and fix the problem soon.

Depends on your definition of "vast", but yeah, it's about 600 million records taking up ~50GB (with indices). Of that, however, most of the data is replicating stuff already known from PrimeNet. The "important" bit is the TF effort above 1000M, which is not duplicated on any other server (but is backed up monthly to my home systems). Once the server is back online, if there's interest I can quite easily make this data available to anyone who might be interested (at a public URL, there's no secret data here).[/QUOTE]

[U]Q[/U]: Does the server at the remote location perform backups more frequently?

50GB. That is not as much as I would have thought. My external HD would easily accommodate that. I wouldn't mind being able to peruse your data, once in a while, just to see what it looks like.

I just checked. Your site is back up and running. :smile:

James Heinrich 2018-04-14 16:51

Yes, the site is back up now and should stay up, hopefully. There may be a reboot or two later today (8h from now) as the RAID gets rebuilt, but otherwise it should be back online for good, running on new hardware.

GP2 2018-04-14 17:11

[QUOTE=James Heinrich;485267]The "important" bit is the TF effort above 1000M, which is not duplicated on any other server (but is backed up monthly to my home systems).[/QUOTE]

Disk storage is cheap. PrimeNet could start storing this data too.

James Heinrich 2018-04-14 18:26

SQL dumps of factor tables are now available at:
<redacted> (~3GB)

TF limit for all exponents now lives at:
<redacted> (~600MB)

I've set it to regenerate these files early in the morning of the 15th of each month.


edit: (2019-Jun-25) the collection of data export downloads are now available at [url]https://www.mersenne.ca/export/[/url]

petrw1 2018-04-15 04:19

[QUOTE=James Heinrich;485300]Yes, the site is back up now and should stay up, hopefully. There may be a reboot or two later today (8h from now) as the RAID gets rebuilt, but otherwise it should be back online for good, running on new hardware.[/QUOTE]

Can I assume we will start seeing current data again in a day or two once RAID is rebuilt?

James Heinrich 2018-04-15 04:27

Data will catch up automatically, yes. I'll check in the morning to make sure it's all up to date, otherwise I'll run the missing days manually.

ixfd64 2018-04-15 04:32

I've noticed that the Brent-Suyama factor page is working again too!

James Heinrich 2018-04-15 15:29

It had caught up two of the missing four days by itself, I gave it a push and updated the other two days, so I think it's all current now. Let me know if you find any missing data.

henryzz 2018-04-15 17:02

[QUOTE=GP2;485301]Disk storage is cheap. PrimeNet could start storing this data too.[/QUOTE]

Hosting bandwidth can be more expensive though.

James Heinrich 2018-04-15 17:25

I usually see it the other way -- you have to pay for additional storage, whereas often [i]bandwidth[/i] (e.g. 250Mbps) is fixed but data transfer is uncapped.
Not that the demand for factor tables should be tremendously high, at least I don't think so.

henryzz 2018-04-15 19:55

[QUOTE=James Heinrich;485365]I usually see it the other way -- you have to pay for additional storage, whereas often [i]bandwidth[/i] (e.g. 250Mbps) is fixed but data transfer is uncapped.
Not that the demand for factor tables should be tremendously high, at least I don't think so.[/QUOTE]

The forum has a data transfer cap. Currently it is a non issue but it was before we upgraded plan.

snme2pm1 2018-04-16 11:37

I recollect that Aaron has provisioned a greater server then ever before, and anticipated it to be under-utilised.
I also recall his long ago comments about lots of spinners! Lovely!
Is not some kind of amalgamation of forces due?
I also believe that Aaron provisioned matters such that the primenet sever was a virtual machine.
As far as I am aware, James is a Linux inclined bloke, due to past head bash emoticons in relation to Widows facilities regarding database nuisances.
So would it be not decent to create a minor related linux virtual machine for James to configure?
Thereby more forces would be combined.

Uncwilly 2018-04-16 14:14

I am of the type that thinks that there is value to having the data and websites spread across multiple different companies resources. What if the company that host PrimeNet has a court action or bankruptcy. Or the data center has a natural disaster or disgruntled employee with plan to do as much damage as possible.

Mirroring the data and custom software that makes up GIMPS should be on at least 3 physical devices, at least 2 wide spread locations, and on different hardware platforms (with at least 2 different storage media [a couple of thumb drives in addition to the SSD/HDD's at the server].

Mersenne.ca, PrimeNet, MF, and GPU72 are all on different companies hardware. Mirroring all datasets and code on to each makes sense. Syncing once a week from the mother site should not max anyone's bandwidth.

Madpoo 2018-04-18 04:51

[QUOTE=snme2pm1;485439]I recollect that Aaron has provisioned a greater server then ever before, and anticipated it to be under-utilised.
I also recall his long ago comments about lots of spinners! Lovely!
Is not some kind of amalgamation of forces due?
I also believe that Aaron provisioned matters such that the primenet sever was a virtual machine.
As far as I am aware, James is a Linux inclined bloke, due to past head bash emoticons in relation to Widows facilities regarding database nuisances.
So would it be not decent to create a minor related linux virtual machine for James to configure?
Thereby more forces would be combined.[/QUOTE]

Reading these posts, the thought did occur to me that maybe a consolidation would be possible. The last time I updated the hardware (when the server was physically moved from Seattle to Dallas) it gained even more CPU/memory/drive space to the point where I'm able to clone Primenet entirely and run a backup copy for testing. We used it to test some of the PRP code and database changes before it went live.

Ultimately I'd defer to George and James on whether either of them would actually want to, but technically it seems possible. The colo provider does give us a good amount of bandwidth and I'm pretty sure we have plenty to spare on a monthly basic. As far as CPU goes, Primenet typically doesn't use much although sometimes the SQL queries may spike it for a bit during more complicated maintenance, but otherwise not too bad.

[QUOTE=Uncwilly;485449]I am of the type that thinks that there is value to having the data and websites spread across multiple different companies resources. What if the company that host PrimeNet has a court action or bankruptcy. Or the data center has a natural disaster or disgruntled employee with plan to do as much damage as possible.
...[/QUOTE]

Speaking just for Primenet, it uses a cloud backup provider with weekly full SQL backups and I think the incremental SQL backups are every 4 hours or something like that (haven't checked lately). The cloud provider I just switched to has an option to backup the entire virtual machine, and I do have the space for it on the account, so maybe I'd look at doing that as well on some quarterly basis. Even though all the data/code is backed up, it would still take some time to rebuild the VM and get it all setup and running so maybe having a full VM snapshot would be a nice-to-have.

Uncwilly 2018-04-18 06:03

[QUOTE=Madpoo;485610]Even though all the data/code is backed up, it would still take some time to rebuild the VM and get it all setup and running so maybe having a full VM snapshot would be a nice-to-have.[/QUOTE]James and Chris have some code that should be preserved. There are some other odds and ends around, some loosely associated with GIMPS and some independent, that would be good to have duplicates (like some of Luigi's things, all of the GPU code sets, a quarterly snap shot of mlucas, etc.)

I know PrimeNet is well set up, but considering that none of the data are on the wayback machine, a well handled second back scheme (mirroring on a different company's platform) might still be wise.

LaurV 2018-04-19 05:49

If external storage space needed, I can offer a corner on my shelves too. If you provide a link, I can download and store what I found there (possibly periodically). 50G is not much, can keep (say) the last 4-5 generations in a 250-300G dedicated backpack...

(edit: assuming it is zipped/encrypted already, otherwise the space to store it will be lower when zipped by myself)

James Heinrich 2018-04-19 10:23

See post #40 for download links, ~3.5GB zipped.

snme2pm1 2018-04-20 13:37

I recall sometime ago there was occasion to speak about the Bus Factor, in regard to PrimeNet.
Well I wonder how that might play into the way that people think in near discussion for mersenne.ca.
Having resources distinctly mirrored, might offer a distinct advantage in safety terms, and I would prefer that James doesn't fall under a bus any time soon (they exhibit too many tons).

kladner 2018-04-20 14:39

My first thought was to wonder which (electronic) bus was in question. I realize now that it is the Bus of Unexpected Fatality. :sad:

retina 2018-04-20 14:44

[QUOTE=kladner;485773]My first thought was to wonder which (electronic) bus was in question. I realize now that it is the Bus of Unexpected Fatality. :sad:[/QUOTE]I think it is the USB. Not the Universal Serial Bus. The Unseen School Bus.

VictordeHolland 2018-04-20 14:55

[QUOTE=James Heinrich;485692]See post #40 for download links, ~3.5GB zipped.[/QUOTE]
I tried downloading the 3GB zipped file, but it stops downloading after about 5-10 minutes (<1GB). I've retried it a couple of times now, without success. Is it something on my side or is the server capping/aborting the transfer?
(downloading 10GB+ zipped files with relations from the NFS server works fine)

The 600MB file download works fine though.

kladner 2018-04-20 15:02

[QUOTE=retina;485774]I think it is the USB. Not the Universal Serial Bus. The Unseen School Bus.[/QUOTE]
LOL!

James Heinrich 2018-04-20 15:51

[QUOTE=VictordeHolland;485777]I tried downloading the 3GB zipped file, but it stops downloading after about 5-10 minutes (<1GB). I've retried it a couple of times now, without success. Is it something on my side or is the server capping/aborting the transfer?[/QUOTE]Nothing intentionally stopping the transfer. If it helps, I've created a .torrent version for each (auto-updated each week) with the http server as webseed
[url]https://www.mersenne.ca/known_factors.sql.gz.torrent[/url]
[url]https://www.mersenne.ca/prime_numbers.sql.gz.torrent[/url]

Mark Rose 2018-04-20 16:09

I was able to download full copies fine yesterday. They decompressed without issue.

[QUOTE=James Heinrich;485792]Nothing intentionally stopping the transfer. If it helps, I've created a .torrent version for each (auto-updated each week) with the http server as webseed
[url]https://www.mersenne.ca/known_factors.sql.gz.torrent[/url]
[url]https://www.mersenne.ca/prime_numbers.sql.gz.torrent[/url][/QUOTE]

Are the dumps being updated weekly now?

What time and day of week (UTC?) are the torrents being updated? I'll cron downloading them to seed them automatically.

chalsall 2018-04-20 16:13

[QUOTE=James Heinrich;485792]If it helps, I've created a .torrent version for each (auto-updated each week) with the http server as webseed[/QUOTE]

I just added these to my rtorrent client, and I'm not seeing any seeders. Are you sure your server has the appropriate ports open?

James Heinrich 2018-04-20 16:52

[QUOTE=Mark Rose;485793]Are the dumps being updated weekly now? What time and day of week (UTC?) are the torrents being updated? I'll cron downloading them to seed them automatically.[/QUOTE]It was monthly, but I can make it weekly. Export now starts 9pm (EST/EDT) every Sat and takes ~15 mins. To be safe (4-5h timezone, an hour for export), you can cron for mirror at/after 0300h every Sunday.

James Heinrich 2018-04-20 16:53

[QUOTE=chalsall;485794]I just added these to my rtorrent client, and I'm not seeing any seeders. Are you sure your server has the appropriate ports open?[/QUOTE]I tried it in uTorrent and it worked fine. All that needs to be open is port 80.

Mark Rose 2018-04-20 16:53

[QUOTE=chalsall;485794]I just added these to my rtorrent client, and I'm not seeing any seeders. Are you sure your server has the appropriate ports open?[/QUOTE]

The web seeding is working fine for me at work. I'll configure rtorrent at home tonight.

Mark Rose 2018-04-20 16:56

[QUOTE=James Heinrich;485801]It was monthly, but I can make it weekly. Export now starts 9pm (EST/EDT) every Sat and takes ~15 mins. To be safe (4-5h timezone, an hour for export), you can cron for mirror at/after 0300h every Sunday.[/QUOTE]

Alright, I'll schedule weekly based on that time.

VictordeHolland 2018-04-20 18:19

The torrents work fine, thanks for making this all possible :bow:.

chalsall 2018-04-20 18:49

[QUOTE=VictordeHolland;485809]The torrents work fine, thanks for making this all possible :bow:.[/QUOTE]

Weird. My rtorrent client still isn't seeing any seeders, yet can download CentOS images just fine.

VictordeHolland 2018-04-20 19:32

[QUOTE=chalsall;485813]Weird. My rtorrent client still isn't seeing any seeders, yet can download CentOS images just fine.[/QUOTE]
What trackers are you getting updates from?

These are two trackers that I got announcements from in the last 30 min:
udp://tracker.opentrackr.org:1337/announce
udp://tracker.open-internet.nl:6969/announce

ixfd64 2018-04-20 20:57

[QUOTE=Uncwilly;485449]I am of the type that thinks that there is value to having the data and websites spread across multiple different companies resources. What if the company that host PrimeNet has a court action or bankruptcy. Or the data center has a natural disaster or disgruntled employee with plan to do as much damage as possible.[/QUOTE]

I agree. Let's not forget what happened to Seventeen or Bust.

snme2pm1 2018-04-21 03:38

[QUOTE=Madpoo;485610]As far as CPU goes, Primenet typically doesn't use much although sometimes the SQL queries may spike it for a bit during more complicated maintenance, but otherwise not too bad.
[/QUOTE]

I don't spend my days immersed with server and VM issues, but I would have thought that a created minor VM would be only utilising a nominated CPU core of the vast number of available CPU cores within the server metal, and prescribed disk and memory and network resources.
So I was unclear about why occasional mersenne.org busy behavior would be significant.
I guess that disk storage contention could be a greater issue for debate, but then perhaps even that is perhaps largely negligible for comment with the huge number of spinners that Aaron has likely incorporated.

GP2 2018-10-01 06:03

If anyone is using the methods posted here to mirror Mersenne.ca's database, especially for the 1G to 10G range which is beyond Primenet's range, please take the time to purge this composite factor from your mirror:

Composite factor of [URL="http://www.mersenne.ca/exponent/3321930431"]M3,321,930,431[/URL] is 2118762583699087568351 = 6643860863 * 318905321377

This composite factor somehow made it (back?) into the Mersenne.ca database even though [URL="https://www.mersenneforum.org/showthread.php?p=289310&postcount=97"]James himself posted about it back in 2012[/URL], which makes me suspect that it was somehow added back in, in some update from some other source at some point.

Note: this is the only composite factor in the Mersenne.ca database.

Possibly by the time you read this, James may have already removed it.

James Heinrich 2018-10-01 11:43

[QUOTE=GP2;497169]Composite factor of [URL="http://www.mersenne.ca/exponent/3321930431"]M3,321,930,431[/URL] is 2118762583699087568351 = 6643860863 * 318905321377
Possibly by the time you read this, James may have already removed it.[/QUOTE]I have no idea how or when it got back in. I have re-removed it.

srow7 2018-10-31 03:06

stats not getting updated

Data as of 2018-10-28T23:59:00+00:00

[url]https://www.mersenne.ca/status/[/url]

James Heinrich 2018-10-31 13:31

It is known. I'm working on it. Should be fixed in an hour or two.

edit: finished updating.

alpertron 2018-12-03 13:55

I found that in page [url]https://www.mersenne.ca/exponent/2535283[/url] , the fifth "prime" factor is the product of the third and fourth.

petrw1 2018-12-03 14:09

I like the menu change
 
More room

Thx

James Heinrich 2018-12-03 17:00

[QUOTE=alpertron;501575]I found that in page [url]https://www.mersenne.ca/exponent/2535283[/url] , the fifth "prime" factor is the product of the third and fourth.[/QUOTE]Indeed. I thought I had found and fixed the codepath that was letting composite factors sneak in... but apparently something's still not quite right :unsure:
I'll have to have another review of how these composite factors keep sneaking into the data. Thanks for letting me know.

edit: I've found and squished another typo that was (occasionally) letting composites through. I hope that's the last of the bugs and no more composites sneak in. Let me know if you can prove me wrong.

lycorn 2018-12-30 09:22

Any chance of updating the number of primes found so our latest achievement shows up on the table?

James Heinrich 2018-12-30 16:07

[QUOTE=lycorn;504348]Any chance of updating the number of primes found so our latest achievement shows up on the table?[/QUOTE][i]monst[/i] just told me about it being missing from [url]https://www.mersenne.ca/status[/url] and that's fixed now. I assume that's the "table" to which you refer?

lycorn 2018-12-30 22:22

[QUOTE=James Heinrich;504376][i]monst[/i] just told me about it being missing from [url]https://www.mersenne.ca/status[/url] and that's fixed now. I assume that's the "table" to which you refer?[/QUOTE]

It is, indeed. Thank you.

James Heinrich 2019-01-02 18:23

I just recently monkeyed with the database structure so it's possible I broke things. If you spot anything amiss, please let me know (either post here or email me directly).

GP2 2019-01-03 00:25

[QUOTE=James Heinrich;504682]I just recently monkeyed with the database structure so it's possible I broke things. If you spot anything amiss, please let me know (either post here or email me directly).[/QUOTE]

I guess this will affect the exported SQL dumps? Do the files available for download reflect this change yet?

By the way, there were 38 factors for hyperlarge exponents (> 10G) in the old global SQL dump files, have these been dropped now?

James Heinrich 2019-01-03 01:57

[QUOTE=GP2;504720]I guess this will affect the exported SQL dumps?[/QUOTE]No, the change only affected the database table storing the GIMPS results imported from PrimeNet. known_factors_* and other exported tables are not affected.

[QUOTE=GP2;504720]By the way, there were 38 factors for hyperlarge exponents (> 10G) in the old global SQL dump files, have these been dropped now?[/QUOTE]No, they're still there, I just hadn't thought to export them. There's a link at the bottom of the export page now.

Chuck 2019-01-09 14:00

I like the new change
 
I like the change in the status display which shows exactly where current assignments are, especially with so many being reserved at a time.

James Heinrich 2019-01-09 14:13

I find it helpful too. Note that these are not realtime results, the assignment count/distribution is only updated once per day, but it's a guide at least.

ramgeis 2019-01-09 17:14

[QUOTE=James Heinrich;505404]I find it helpful too. Note that these are not realtime results, the assignment count/distribution is only updated once per day, but it's a guide at least.[/QUOTE]
A nice feature indeed.


By the way, in the assignment request form on that page the "Min current TF" value seems to have no effect. Whatever I enter there, the value seems to be replaced by 65.

James Heinrich 2019-01-09 18:08

[QUOTE=ramgeis;505420]in the assignment request form on that page the "Min current TF" value seems to have no effect.[/QUOTE]Thanks. Please try again, it should work better now.

ramgeis 2019-01-09 23:25

[QUOTE=James Heinrich;505430]Thanks. Please try again, it should work better now.[/QUOTE]
it still doesn't... I requested assignments from 68 to 69 and got assignments from 67 to 69

James Heinrich 2019-01-09 23:28

[QUOTE=ramgeis;505474]it still doesn't... I requested assignments from 68 to 69 and got assignments from 67 to 69[/QUOTE]Sorry, SPE... I fixed it locally and forgot to upload the changes :redface:

ramgeis 2019-01-10 01:24

[QUOTE=James Heinrich;505475]Sorry, SPE... I fixed it locally and forgot to upload the changes :redface:[/QUOTE]
now it works, thanks...
one more thing: the link in the cells of the table which automatically fills the form contains min/max values which are both one to high

James Heinrich 2019-01-10 09:14

[QUOTE=ramgeis;505488]the link in the cells of the table which automatically fills the form contains min/max values which are both one too high[/QUOTE]Also fixed, thanks.

alpertron 2019-01-16 23:16

[url]https://www.mersenne.ca/exponent/2568793[/url]

The fourth factor is the product of the second and third. Probably this problem occurs because the Prime95 application executing on my computer submitted the product of the second and third prime numbers while running the P-1 algorithm.

petrw1 2019-01-17 17:58

From James via email
 
[QUOTE] I'm aware of problems but on vacation until 20 Jan, I have email-only access, can't post on forum and can't fix my server. Thanks :)[/QUOTE] .

James Heinrich 2019-01-21 02:46

[QUOTE=alpertron;506158][url]https://www.mersenne.ca/exponent/2568793[/url]
The fourth factor is the product of the second and third. Probably this problem occurs because the Prime95 application executing on my computer submitted the product of the second and third prime numbers while running the P-1 algorithm.[/QUOTE]I have removed the offending factor, but I am not yet certain how or why it was allowed into the database in the first place. Further investigation is required. :mad:

nomead 2019-01-21 05:36

Small cosmetic thing. On the tf1G.php page, the counts in the Assigned column can go negative. This seems to happen when an exponent is reserved for factoring more than one bit depth in one go, say 67 to 69, and then the results are reported in two parts, first 67 to 68, and then 68 to 69.

Anyway all is correct after the daily run so it's just a temporary nuisance.

James Heinrich 2019-01-21 11:26

Yes, that can happen as you describe. For speed of display the counts shown on the page are from a cache table that stores the numbers assigned and available, and when results are reported and accepted the count is updated. The side effect is as you describe: if a result is submitted and accepted more times than the exponent is assigned (multiple bit levels reported for a single assignment, or results reported for unassigned exponents) then the count can temporarily end up negative. The data is rebuilt every night and currently takes about 16 minutes to rebuild.

Syntony 2019-03-23 00:29

Exponent status report "Needs double-check" erroneously
 
Recently the exponent status reports seem to give "needs double-check" despite matching LL residues... e,g. [URL]https://www.mersenne.org/report_exponent/?exp_lo=46851641&full=1[/URL] and [URL]https://www.mersenne.ca/exponent/46851641[/URL]


All times are UTC. The time now is 03:50.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.