Grrr... 1and1 'auto-block' has blocked my home IP address

Pages: << < (3/3)

Johnmac:

I realize that this is an older thread, but I am having the same issue with justhost.com after they moved me over to a new server. I am using Fling to upload my data at 5 minute intervals. The error message that I get is that the password is no good, but I know it is correct as the data has been uploading just fine.

They whitelisted my IP, but that does not do any good, if you have another (what they call an attack from the IP) it blocks your IP again. After they reset it, Fling will work for 2 or 3 days then it is denied again. I have changed my upload times from 5 minutes to 10 minutes after this latest block. Will see if that helps.

Is it possible that Fling could send the wrong password after a few days of operation? I rebooted today, lets see if that helps. I did read elsewhere that Fling had some leaks or handle issues. This problem may have started after updating Fling, but I am not sure.

saratogaWX:

I think that more hosters are now checking for concurrent connection limits like 1and1 is doing, so this may become a more common problem, especially for enthusiasts who run multiple weather software packages (like me).

I use Weather-Display (with some stuff from WeatherLink) to run my main site, but also have multiple weather software running (mostly for testing with the template sets).  So my main site has:

Weather-Display (1 connection for clientrawrealtime uploads, 1 connection for main WD ftp)
GRLevel3 (1 connection)
WASP2 (two instances, 1 connection each)
Cumulus (1 connection for realtime FTP, 1 for main files FTP)

Fling supports (with 1 connection) uploads from WeatherLink, VWS (including wflash realtimes), MeteoHub, WeatherSnoop, WeatherCat as needed.

The issue with Fling was a steady consumption of non-pagable memory which used to cause a system crash every 3 days or so.  I couldn't find a fix from the vendor, so I did a small Perl script to run via Windows Scheduler to check the memory for the Fling task every 5 minutes, and kill the Fling system task when it exceeded 16Mb.  It auto restarts after the 'kill' so no data is lost, but the non-pagable memory is freed up by the kill :)   Now my system stays up until reboot after the Microsoft 'black tuesday' patches are installed each month.

Don't know if this will help your issue, but I offer the script I'm using for Fling kill.

You'll need ActiveState's Perl (http://www.activestate.com/activeperl/downloads ) installed,
Microsoft's pslist and pskill commands from the PsTools set (http://technet.microsoft.com/en-us/sysinternals/bb896649 )
And to set up a Scheduled Task to run the zap-fling.pl at your desired interval.

zap-fling.pl: Code:
--
#!/usr/bin/perl
#
# use the pstools to find the current NP memory for fling, and optionally
# use pskill to zap it to release the held memory
#
# K. True - 30-Oct-2011
#
# Run via Windows Scheduler or Linux cron .. all output is to the log file(s)
#
# ---- configurable settings ----
$maxMem = 16384;  # amount in KB that we'll tolerate

$logsDir = "./fling-logs/"; # place to store the YYYYMMDD.txt

# ---- end of configurable settings --
#
$|=1; # no buffering of output

 $cur_time = time();
 @months = ("Jan","Feb","Mar","Apr","May","Jun","Jul","Aug","Sep","Oct","Nov","Dec");

 $log_filename = ISO_datestamp($cur_time) . ".txt";
 $nice_date = nice_datestamp($cur_time);

 open (OUT,">>${logsDir}${log_filename}");

open(PS,"pslist -m fling |") || die "..unable to run pslist $!\n";

while (<PS>)
{
  #print $_;
  $rec = $_;
  ($name,$PID,$VM,$WS,$Priv,$PrivPK,$Faults,$NonP,$Page) = split(/\s+/);
  next unless $Page =~ m/\d+/;
  #print STDOUT "$name\t$PID\t$NonP\n";
  print OUT "$nice_date\t$rec";
  print STDOUT "$nice_date\t$rec";
 
  if($NonP >= $maxMem) {
    print OUT "$nice_date\t$name PID=$PID pskill for $NonP > $maxMem KB\n";
    print STDOUT "$nice_date\t$name PID=$PID pskill for $NonP > $maxMem KB\n";
open(PK,"pskill $PID |") || die "..unable to run pslist $!\n";
while (<PK>) {
   print OUT "$nice_date\t$_";
   print STDOUT "$nice_date\t$_";
 
}
    print OUT "$nice_date\t-------------------------------------------\n";
    print STDOUT "$nice_date\t-------------------------------------------\n";

  }
}

# ------  end of main program -----

sub nice_datestamp {
    my $d = shift;
   
    my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) = localtime($d);
    my $nicedate = sprintf(
        "%02d\-%s\-%04d %02d\:%02d\:%02d",
        $mday,$months[$mon],$year+1900,$hour,$min,$sec);
    return("$nicedate");

}

sub ISO_datestamp {
    my $d = shift;
   
    my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) = localtime($d);
    my $nicedate = sprintf("%04d%02d%02d",$year+1900,$mon+1,$mday);
    return("$nicedate");

}

--

You'll need a ./fling-logs/ directory to store the YYYYMMDD.txt files created by the script which contains the log of the activities so you'll see when fling gets zapped.  I've attached a sample from my system.

Hope this helps...

Best regards,
Ken

Johnmac:

Ken,

Thanks for the suggestions. For now I have added the Fling program to my "Startwatch" program that I have been using for some other programs. It gives the option of killing the Fling.exe file if the memory gets too high. Based on your comments, I set that level at 16mb, but feel that may be too high based on the usage the Startwatch program is reporting. I will watch it and see if the memory starts creeping up. Currently it is less than 2mb.

Thanks for the comments to help solve my problem.

weatherc:

Quote
--
I think that more hosters are now checking for concurrent connection limits like 1and1 is doing, so this may become a more common problem, especially for enthusiasts who run multiple weather software packages (like me).
--

It is. I have seen more and more wx-sites kicked out due to this and are slowly being almost a problem for us wx-sites. Not actually for me as i have own server but for those on webhotels. Seems webhosts trys to get more sites to fit on the boxes with tighter limits or something.

When i installed cPanel on the dedicated server i use i started with default settings for FTP. On same server are a handful wxsites. It tookn't long time before the FTP-limit was reached, and my FTP'ers are only WD and NSLog + a scheduled batscript every now and then. Plus then my "own connections" sometimes Filezilla,SSH and browser.
I don't remember what the deafult setting was but below 10 it was and only after set it to near 20 the problem went away.

I think there may also sometimes be some connections what maybe not are closed properly and hanging around for some time what increase the amount of connections.

So, yes, its pretty easy to reach the roof of those if not webhoster has increased them....

Henkka

DaculaWeather:

You know I had the problem not long ago when they told me I had 186,000 connections on my server!!! IXWebHosting shut me down for about 6 hours. I had to raise enough hell to get them to turn me back on. And of course when I asked them what I could have done about it, they said nothing. Thanks. So shut me down and not tell me, for something I have no control over. Makes perfect sense.

Pages: << < (3/3)