OFFICIAL "SERVER PROBLEMS" THREAD

[ewmayer]

Please post any information about problems you're having with the PrimeNet server here.

Note to the Moderators: Please clear posts out of this thread as the relevant issues are resolved, or the problems simply go away due to server reboot and such. Posts that prove to address specific problems that are not merely due to transient glitches or server-is-down-ness and where a server-side fix is needed should be moved into separate threads. Thanks.

[Old man PrimeNet]

For immediate communications please email me the particulars of any issue at primenet@mersenne.org. I respond to email significantly earlier than forum posts.

For participants using or upgrading from Prime95/MPrime v24 or earlier (installed before October 2008), you will need to make a new PrimeNet account and then link your old account to the new account to transfer any accumulated work credits -- please start here: http://www.mersenne.org/gettingstarted/ .

For participants picking your own assignments using the PrimeNet APIs directly for bulk advance TF in particular:
(a) Note that this activity is not encouraged and there's a large probability that in the near future PrimeNet will either restrict or block this kind of activity completely. If your IP gets blocked it means you are not paying attention to your activity's impact on the PrimeNet server's response times and instead it has my attention, and not the good kind.
(b) If your computers are using the native communications and work-type selection features of Prime95 and MPrime, PrimeNet will typically automatically assign them the most useful work for GIMPS' objectives. The worktodo files maintained by the programs are designed to be let alone and queue the work for that computer only, not used as a workload redistribution point using other mechanisms.
(c) If you insist upon doing these things, consider using the Manual Results web forms, which are the intended method for requesting and submitting large numbers of work units the server would not normally assign to your computers.
(d) Lately there has been a noticeable increase in advance, shallow (e.g. rapid) TF unit workload, which is perhaps the least useful work and generates the most network traffic, and is why the server usually will not assign these in significant counts. While we are grateful for all useful contributed computational work, the GIMPS PrimeNet server has and will always have finite resources, and anyone insisting upon doing high-volume race-ahead work of any kind is required to exercise the necessary programming skill and discipline to continuously monitor and dynamically react to the impact of their web API traffic so as to not bog down the entire GIMPS project in pursuit of their personal objectives, ignorantly or otherwise. If response times lengthen significantly or the API repeatedly times-out, your program may be transacting too frequently and should use a back-off rate timing scheme.

[Old man PrimeNet]

Old IPs 71.6.220.162, .163
New IPs 66.181.10.42, .43

Expected total server downtime about 90 minutes 2013-08-17 12:00 PDT / 19:00 UTC

[Uncwilly]

I think that either the server or the database itself is having issues.

I have been checking the status of the 332M range (specifically those at or above the 100M digit level.) I check pointed the counts for the exponents at various bit levels. They have not changed since 8/20. My machines have turned in at least 9 results in that range (more if you count exponents that one machine turned in that got miscredited, the server said that the AID belonged to someone else.)

[chalsall]

Originally Posted by Uncwilly

I think that either the server or the database itself is having issues.

Care to give explicit examples? Exponents and AIDs would be helpful.

Edit: Please send directly to George or Scott, not publicly.

[Uncwilly]

Originally Posted by chalsall

Care to give explicit examples? Exponents and AIDs would be helpful.

Edit: Please send directly to George or Scott, not publicly.

Looks like somebody kicked it. They are now updated.

I will get the AID's once I get near the machine in question.

[Uncwilly]

Originally Posted by Uncwilly

I think that either the server or the database itself is having issues.

Originally Posted by Uncwilly

Looks like somebody kicked it. They are now updated.

Note to self, turn automatic calculation back on in Excel.

[ixfd64]

I'm getting the following error when I try to access the GIMPS website:

pnErrorResult=3 pnErrorDetail=Database unavailable ==END==

[kracker]

Originally Posted by ixfd64

I'm getting the following error when I try to access the GIMPS website:

pnErrorResult=3 pnErrorDetail=Database unavailable ==END==

Looks like it's back up!

[kladner]

I uploaded successfully ~1730 UTC.

[chalsall]

Originally Posted by ixfd64

I'm getting the following error when I try to access the GIMPS website:

pnErrorResult=3 pnErrorDetail=Database unavailable ==END==

This is just what one has to accept and manage when dealing with Primenet. The Database is not always available....