|
|
|
|
 |
|
May 13, 2008 |
|
|
 |
|||||
|
Home |
About This Site |
Contact |
Access Statistics |
|||||
 |
|
|
|
|
|
 Email Me Résumé HTMLPerlSETI Remote Controls Home Theater For Sale Public Links Private Links        |
HTMLPerlSETI
09/02/2002 version 0.17
SETI-STATS : HTMLPerlSETI Note 1 If you have HTMLPerlSETI configured to get stats from a remote machine, SETI::Stats assumes that the login account (that rsh or ssh logs in as) is a Bourne flavored shell (for example: sh, bash, ksh, zsh). If it is a C shell flavor (csh, tcsh), you'll need to edit line 76 of Stats.pm and delete the "2>/dev/null" so the line looks like: open(IN, "$self->{rsh} $self->{host} \"cat $filename\"|"); The "2>/dev/null" notation means to redirect stderr to /dev/null. C shell does not understand this (or have syntax to redirect stdout and stderr separately). Note 2 If you have HTMLPerlSETI configured to use ssh or ssh2 to get the stats from a remote machine, and HTMLPerlSETI hangs when it ssh's to the remote machine, (and you have everything configured correctly) try this change in the SETI::Stats Stats.pm file. Change "cat $filename" to "sed -n '1,/^end_seti_header/p' $filename" on line 76 of Stats.pm, so the line looks like: open(IN, "$self->{rsh} $self->{host} \"sed -n '1,/^end_seti_header/p' $filename 2>/dev/null\"|"); Thanks to Jean-Sebastien Morisset for this sed fix. An explanation for technies, if any of you are interested: This is how reading the work_unit.sah file in Stats.pm is supposed to work. There is a while loop that reads the output of the cat $filename command. On the work_unit.sah file there is about 50 lines of data (stats) that we want and then there is the actual work unit. There is a line that starts off with "end_seti_header" to signify the end of the stats and that the next line begins the work unit. In the Stats.pm code it looks for the end_seti_header and when it finds it, it stops reading and breaks out of the while loop and then closes the file (actually it closes the cat pipe). I don't know why but, for some users, it exits the while loop exactly like it's supposed to, but then it hangs on the close. I'm baffled as to why this only happens for a few users. Perhaps it's caused by how certain machines buffer the pipe via ssh. If you have any ideas, feel free to send me an email. |
|||||||||
 |
|
|||||||||
 |
