Re: [PLUG] Job running in background

[Erik Lane]

Just a quick note - there is practically no learning curve to screen.

You can start it with just 'screen' then run whatever commands you want,
and then you can disconnect it with Ctrl-a d. (Ctrl-a pushed together, then
release both, then press d by itself.) Then 'screen -r' will reconnect to
it to check what it's doing.

Of course there are tons more options, but these are really what I mainly
use it for. I love being able to disconnect from the machine while leaving
my processes running. I even run screen from cron sometimes, if it's a
recurring job that I want to be able to check in on while it's running.
(For example, one that might run for days... :)  )

On Thu, Mar 2, 2017 at 10:55 AM, Rich Shepard 
wrote:

> On Thu, 2 Mar 2017, Tom wrote:
>
> > To see if your background process does something with the CPU rather than
> > just hang around waiting for user input for example - you could see the
> > process CPU load and wall clock time - the time should be increasing. The
> > easiest way is to fire: top and see what CPU percentage the process uses
> > and if the TIME+ metrics is increasing.
>
>
> Tom,
>
>The model runs without user input. I ran top and it showed python
> running
> (because the model's written in python). I don't recall what the time
> metrics showed.
>
> > Please note that some processes do not run in background - if you put
> > them in background they sleep until they are in foreground. If that is
> > the case for your process, you need screen, tmux or vnc to keep the
> > process in the foreground. Also some progress indicator - in a log file
> > perhaps would be great to have for long running applications.
>
>It could be that this model does not like running in the background.
> I've
> modified a couple of run parameters and started it again, but in the
> foreground this time. The author tells me that the completion ETA will
> increase until soils are saturated, then decrease drastically. I'm looking
> forward to seeing a decrease in ETA because now the model run is 1.3%
> completed and the ETA is 3 days, 12 hours, and increasing. :-)
>
>I've not before run a hydrologic model over a 25 day period with
> precipitation specified in mm/hr. A learning experience all around.
>
> Thanks,
>
> Rich
> ___
> PLUG mailing list
> PLUG@lists.pdxlinux.org
> http://lists.pdxlinux.org/mailman/listinfo/plug
>
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Web browsers not displaying images

[Chuck Hast]

I had I think it was linkdin got idiot for a bit this morning but it came
back. I am
doing job hunting, and was seeing those weird screens there for a while. It
is
bad enough dealing with Hughesnet that can not seem to get links to some
sites that I am trying to reach, I have a wait a bit and THEN I can get
there. I
hope to soon find a job, and get moved to some place where we do not have
to live with Hughesnet, whattapieaceofcrap...


On Thu, Mar 2, 2017 at 9:05 AM, Rich Shepard 
wrote:

> On Thu, 2 Mar 2017, Chuck Hast wrote:
>
> > I saw that also, did the usual close down, no joy, reboot, no joy, then I
> > tested from another machine and got the same thing so figured it was "out
> > yonder somewhere". I still have a couple that come up like that but most
> > of them are back to normal. This was on both FF and Chrome. Looks like WU
> > is back. I see it all correctly now.
>
>Thanks, Chuck. Still no improvements here. Guess I'll wait and see if
> things get better on the desktop. Can use firefox on the laptop if
> necessary.
>
> Thanks,
>
> Rich
> ___
> PLUG mailing list
> PLUG@lists.pdxlinux.org
> http://lists.pdxlinux.org/mailman/listinfo/plug
>



-- 

Chuck Hast  -- KP4DJT --
Glass, five thousand years of history and getting better.
The only container material that the USDA gives blanket approval on.
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] monitoring MAGNITUDE of internet activity - how?

[Keith Lofstrom]

On Wed, Mar 01, 2017 at 04:33:57PM -0600, Richard Owlett wrote:
> Is there a tool which can be
>  "turned on"  at time t0
>  "turned off" at time t1
> which will report the number of
>   "uploaded"   bytes
>   "downloaded" bytes
> in that interval?

For your purposes, perhaps you pipe a standard interface
reporting tool into a text file that you can edit into
.cvs format for a spreadsheet. 

However, sometimes the easiest way to do stuff is to learn
a scripting language and automate a task.  The good thing
about scripting languages is that you can always look at
the program to remind yourself how it works.

A decade ago, I wrote a cheesy perl hack to wrap around
"ifconfig" to look at total bandwidth used, then modified
it a year ago to look at specific interface usage.  It is
very dependent on the behavior of /sbin/ifconfig for an
old distro, but it is easy to modify.  It would be even
easier if "standard" tool maintainers were more
intelligent and responsible and less "creative".

Keith


--
#!/usr/bin/perl
#  ifbr60 - average traffic rate for 60 seconds
#  depends on specific behavior of /sbin/ifconfig
#  V0.1.1  Keith Lofstrom   KLIC   2016 Jan 30

my $delay = 60 ;

use bigint ;  # the byte counts can be very large
my $ifc;  # interface name
my $encap  ;  # type of encapsulation
my $mac;  # mac (hardware) address
my $updn   ;  # interface up or down?
my $inet   ;  # ipv4 internet address
my $txrx   ;  # sum of transmit and receive bytes
my $err;  # sum of transmit and receive errors

my $ifcNum = scalar(@ARGV);
my @r, @t;

for my $pass (0..1) {
  my $ifcnt = -1 ;
  foreach my $port (@ARGV) {
$ifcnt  += 1 ;  # start at 0
open (IFOUT, "/sbin/ifconfig |" );
  while () {
if( /(\w+)\s+Link encap:(\w+)/ ) {
   $ifc   = $1 ;
} elsif( /inet addr:([\d\.]+)/ ) {
   $inet = $1 ;
} elsif( /RX bytes:(\d+).+TX bytes:(\d+)/ ) {
   if( $port eq $ifc ) {
  printf "%4s%14s%16s.r%16s.t\n", $ifc,$inet,$1,$2 ;
  if( $pass == 0 ) {
 $r[$ifcnt] = $1 ;
 $t[$ifcnt] = $2 ;
  } else {
 my $rd = ( $1 - $r[$ifcnt] ) / $delay ;
 my $td = ( $2 - $t[$ifcnt] ) / $delay ;
 printf "%4s%14s%15.3f/s %15.3f/s\n\n", $ifc,$inet,$rd,$td ;
  }
   }
}
  } 
   }
   if( $pass == 0 ) {
  printf("\n");
  sleep $delay;
   }
} 
close IFOUT;


-- 
Keith Lofstrom  kei...@keithl.com
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Job running in background

[Rich Shepard]

On Thu, 2 Mar 2017, Tom wrote:

> To see if your background process does something with the CPU rather than
> just hang around waiting for user input for example - you could see the
> process CPU load and wall clock time - the time should be increasing. The
> easiest way is to fire: top and see what CPU percentage the process uses
> and if the TIME+ metrics is increasing.


Tom,

   The model runs without user input. I ran top and it showed python running
(because the model's written in python). I don't recall what the time
metrics showed.

> Please note that some processes do not run in background - if you put
> them in background they sleep until they are in foreground. If that is
> the case for your process, you need screen, tmux or vnc to keep the
> process in the foreground. Also some progress indicator - in a log file
> perhaps would be great to have for long running applications.

   It could be that this model does not like running in the background. I've
modified a couple of run parameters and started it again, but in the
foreground this time. The author tells me that the completion ETA will
increase until soils are saturated, then decrease drastically. I'm looking
forward to seeing a decrease in ETA because now the model run is 1.3%
completed and the ETA is 3 days, 12 hours, and increasing. :-)

   I've not before run a hydrologic model over a 25 day period with
precipitation specified in mm/hr. A learning experience all around.

Thanks,

Rich
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Job running in background

[Tom]

To see if your background process does something with the CPU rather
than just hang around waiting for user input for example - you could
see the process CPU load and wall clock time - the time should be
increasing.
The easiest way is to fire: top and see what CPU percentage the process
uses and if the TIME+ metrics is increasing.
Please note that some processes do not run in background - if you put
them in background they sleep until they are in foreground. If that is
the case for your process, you need screen, tmux or vnc to keep the
process in the foreground. Also some progress indicator - in a log file
perhaps would be great to have for long running applications.
Tomas
On Thu, 2017-03-02 at 08:38 -0800, Rich Shepard wrote:
> On Thu, 2 Mar 2017, Larry Brigman wrote:
> 
> > If you started it with nohup; it will stay running until it exits
> > on its
> > own or is killed. The output of jobs (based on my reading) is local
> > to the
> > running shell. So each window/shell could result in different
> > output. If
> > the shell closes, the jobs list is removed. Current running jobs
> > would
> > receive a SIGHUP when the shell closes. NOHUP causes the shell to
> > not send
> > the signal, just like if disown job_num was issued.
> 
> Larry,
> 
>I forgot that nohup is local to the starting shell.
> 
> > These background jobs tend to be monitored by a status file.  If
> > you don't
> > have some way to determine how far into the program/data the
> > program has
> > progressed you won't know it has stalled.
> 
>That's true.
> 
> > Putting this in screen or tmux will allow you to come back to the
> > session
> > even if you have totally logged out of the machine. To check the
> > status,
> > login and re-attach to the (still) running session.
> 
>A good lesson. Next time I'll learn how to use screen or tmux
> instead of
> nohup. But, I have a time-step issue to result first. Running in the
> foreground the ETA for the run completion keeps increasing so I need
> to
> learn how to get daily time steps rather than one-second time steps.
> A
> message to the model's author is waiting for a response.
> 
> Thanks very much,
> 
> Rich
> ___
> PLUG mailing list
> PLUG@lists.pdxlinux.org
> http://lists.pdxlinux.org/mailman/listinfo/plug
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Seeking opinion on new century link 1G install

[John Jason Jordan]

On Thu, 2 Mar 2017 00:03:51 -0800
Bill Barry  dijo:

>> Still can't get the desktop to see the internet. :(

>You could at least temporarily set the network up on the desktop to
>use DHCP and see if that fixes the problems. If you later want to
>switch it back to static ip addresses you can use the  settings it
>gets via DHCP as a guide.

Success!

Previously, from advice found on Ubuntu forums I had added the following
lines to /etc/network/interfaces:

auto eth0
iface eth0 inet static
address 192.168.1.146
gateway 192.168.1.1
netmask 255.255.255.0

It didn't seem to make any difference, so this morning I #'d them out.
Then I searched all over for the correct command on 14.04 to restart
the network. I finally found the right command 'sudo service
network-manager restart.' After issuing the command Firefox was able to
go places and Banshee was able to stream internet radio, so all is
right with my world again.

Many thanks to all who helped! 
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Web browsers not displaying images

[Rich Shepard]

On Thu, 2 Mar 2017, Chuck Hast wrote:

> I saw that also, did the usual close down, no joy, reboot, no joy, then I
> tested from another machine and got the same thing so figured it was "out
> yonder somewhere". I still have a couple that come up like that but most
> of them are back to normal. This was on both FF and Chrome. Looks like WU
> is back. I see it all correctly now.

   Thanks, Chuck. Still no improvements here. Guess I'll wait and see if
things get better on the desktop. Can use firefox on the laptop if
necessary.

Thanks,

Rich
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Web browsers not displaying images

[Chuck Hast]

I saw that also, did the usual close down, no joy, reboot, no joy, then I
tested
from another machine and got the same thing so figured it was "out yonder
somewhere". I still have a couple that come up like that but most of them
are
back to normal. This was on both FF and Chrome. Looks like WU is back. I
see it all correctly now.

On Thu, Mar 2, 2017 at 8:41 AM, Rich Shepard 
wrote:

> On Thu, 2 Mar 2017, Larry Brigman wrote:
>
> > I saw a report on the AWS S3 outage. Lots of companies place their static
> > files in S3 storage. This is probably what hit you and a bunch of other
> > sites.
>
> Larry,
>
>I saw that report, too. That was on Tuesday, the last day sites
> displayed
> correctly here. If it's a cache issue (because these sites are still not
> correctly displaying on the desktop, but do on the laptop) that should have
> been cleared during the rebooting. No?
>
>Think I'll manually clear the cache and see if that makes a difference.
>
> Thanks again,
>
> Rich
> ___
> PLUG mailing list
> PLUG@lists.pdxlinux.org
> http://lists.pdxlinux.org/mailman/listinfo/plug
>



-- 

Chuck Hast  -- KP4DJT --
Glass, five thousand years of history and getting better.
The only container material that the USDA gives blanket approval on.
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Web browsers not displaying images

[Rich Shepard]

On Thu, 2 Mar 2017, Rich Shepard wrote:

> Think I'll manually clear the cache and see if that makes a difference.

   Sigh. No joy.

   Other ideas welcome.

Rich
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Web browsers not displaying images

[Rich Shepard]

On Thu, 2 Mar 2017, Larry Brigman wrote:

> I saw a report on the AWS S3 outage. Lots of companies place their static
> files in S3 storage. This is probably what hit you and a bunch of other
> sites.

Larry,

   I saw that report, too. That was on Tuesday, the last day sites displayed
correctly here. If it's a cache issue (because these sites are still not
correctly displaying on the desktop, but do on the laptop) that should have
been cleared during the rebooting. No?

   Think I'll manually clear the cache and see if that makes a difference.

Thanks again,

Rich
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Job running in background

[Rich Shepard]

On Thu, 2 Mar 2017, Larry Brigman wrote:

> If you started it with nohup; it will stay running until it exits on its
> own or is killed. The output of jobs (based on my reading) is local to the
> running shell. So each window/shell could result in different output. If
> the shell closes, the jobs list is removed. Current running jobs would
> receive a SIGHUP when the shell closes. NOHUP causes the shell to not send
> the signal, just like if disown job_num was issued.

Larry,

   I forgot that nohup is local to the starting shell.

> These background jobs tend to be monitored by a status file.  If you don't
> have some way to determine how far into the program/data the program has
> progressed you won't know it has stalled.

   That's true.

> Putting this in screen or tmux will allow you to come back to the session
> even if you have totally logged out of the machine. To check the status,
> login and re-attach to the (still) running session.

   A good lesson. Next time I'll learn how to use screen or tmux instead of
nohup. But, I have a time-step issue to result first. Running in the
foreground the ETA for the run completion keeps increasing so I need to
learn how to get daily time steps rather than one-second time steps. A
message to the model's author is waiting for a response.

Thanks very much,

Rich
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Job running in background

[Larry Brigman]

If you started it with nohup; it will stay running until it exits on its
own or is killed.  The output of jobs (based on my reading) is local to the
running shell.  So each window/shell could result in different output.  If
the shell closes, the jobs list is removed.  Current running jobs would
receive a SIGHUP when the shell closes.  NOHUP causes the shell to not send
the signal, just like if disown job_num was issued.

These background jobs tend to be monitored by a status file.  If you don't
have some way to determine how far into the program/data the program has
progressed you won't know it has stalled.

Putting this in screen or tmux will allow you to come back to the session
even if you have totally logged out of the machine.
To check the status, login and re-attach to the (still) running session.

On Thu, Mar 2, 2017 at 6:22 AM, Rich Shepard 
wrote:

>Knowing that a model would run for a long time I started it last Sunday
> at
> 8:10 am in the background by appending '&' to the command line. Since then
> I
> sporadically look at the status of that process ID and see the status as S,
> which I understand is interruptible sleep.
>
>When I check status with 'jobs -l' there's nothing returned so why does
> it
> continue to display in the process list?
>
> Rich
> ___
> PLUG mailing list
> PLUG@lists.pdxlinux.org
> http://lists.pdxlinux.org/mailman/listinfo/plug
>
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

[PLUG] Web browsers not displaying images

[Rich Shepard]

   Starting yesterday morning firefox-45.7.0 did not display images on
certain web pages, for example, the radar on wunderground.com and all images
on oregonlive.com. When I try to load those pages using
chromium-54.0.2840.100 they won't load at all. Other web sites also have
issues: linkedin.com displays no images and after I log in there's a 'busy'
icon continuously displaying without end; amazon.com lets me log in but not
log out.

   This morning I tested these sites from a different host and they all
displayed and worked as expected. So the problem is on my desktop
server/workstation.

   I rebooted the system but this does not fix the problem. I'm at a loss
where to look for the source of this new issue as I've not before
encountered anything like it.

   I'd like your thoughts on diagnosing the problem so I can resolve it.

Rich

___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

[PLUG] Job running in background

[Rich Shepard]

   Knowing that a model would run for a long time I started it last Sunday at
8:10 am in the background by appending '&' to the command line. Since then I
sporadically look at the status of that process ID and see the status as S,
which I understand is interruptible sleep.

   When I check status with 'jobs -l' there's nothing returned so why does it
continue to display in the process list?

Rich
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] monitoring MAGNITUDE of internet activity - how?

[Richard Owlett]

On 03/01/2017 07:23 PM, King Beowulf wrote:
> On 03/01/2017 02:33 PM, Richard Owlett wrote:
>> *NOTE BENE*
>> Capitalization in subject line semantically important :>
>>
>> I have significant bandwidth constraints.
>> I have been asked to participate in a project requiring significant data
>> transfers.
>> I wish to run a test case to estimate the connectivity "cost".
>> If it is significant, I use Debian Jessie.
>> For the test case the source/destination is of secondary importance.
>>
>> Is there a tool which can be
>>   "turned on"  at time t0
>>   "turned off" at time t1
>> which will report the number of
>>"uploaded"   bytes
>>"downloaded" bytes
>> in that interval?
>>
>> If the tool reports more, no problem -- it will likely be ignored.
>> TIA
>>
>
> I forgot about this one.  Probably easier to use as a shell script:
> http://humdi.net/vnstat/
>

Not fine grained enough for the project which triggered my question.
However, it can likely answer questions that haven't arisen yet. As it's 
in the Debian repository I'll install and give it a test run.




___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] monitoring MAGNITUDE of internet activity - how?

[Richard Owlett]

On 03/01/2017 05:32 PM, Richard Owlett wrote:
> On 03/01/2017 04:44 PM, Robert Citek wrote:
>> Would /proc/net/dev give you the info you are looking for?
>
> I don't know.
> Will investigate when awake.
> Thank you.

A good night's sleep and a cup of coffee (even if decaf) yields a 
*POSITIVE* verdict. It and some simple scripts will allow me record with 
appropriate resolution/granularity different processes which could last 
from minutes to hours. It seems "made for" the problem that prompted my 
question.


>
>
>>
>> $ sudo cat /proc/net/dev
>> Inter-|   Receive|  Transmit
>>  face |bytespackets errs drop fifo frame compressed
>> multicast|bytespackets errs drop fifo colls carrier compressed
>>   eth0: 16839226   26454000 0  0 0
>> 16903410   27268000 0   0  0
>>   eth1:   0   0000 0  0 0
>> 578   7000 0   0  0
>> lo:   0   0000 0  0 0
>>   0   0000 0   0  0
>>
>> Capture at t0 and t1, parse, diff.
>>
>> Regards,
>> - Robert
>>
>> On Wed, Mar 1, 2017 at 2:33 PM, Richard Owlett  wrote:
>>> *NOTE BENE*
>>> Capitalization in subject line semantically important :>
>>>
>>> I have significant bandwidth constraints.
>>> I have been asked to participate in a project requiring significant data
>>> transfers.
>>> I wish to run a test case to estimate the connectivity "cost".
>>> If it is significant, I use Debian Jessie.
>>> For the test case the source/destination is of secondary importance.
>>>
>>> Is there a tool which can be
>>>   "turned on"  at time t0
>>>   "turned off" at time t1
>>> which will report the number of
>>>"uploaded"   bytes
>>>"downloaded" bytes
>>> in that interval?
>>>
>>> If the tool reports more, no problem -- it will likely be ignored.
>>> TIA
>>>
>>> ___
>>> PLUG mailing list
>>> PLUG@lists.pdxlinux.org
>>> http://lists.pdxlinux.org/mailman/listinfo/plug
>> ___
>> PLUG mailing list
>> PLUG@lists.pdxlinux.org
>> http://lists.pdxlinux.org/mailman/listinfo/plug
>>
>
>
> ___
> PLUG mailing list
> PLUG@lists.pdxlinux.org
> http://lists.pdxlinux.org/mailman/listinfo/plug
>


___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

[PLUG] TONIGHT: PLUG General Meeting: UnPLUG

[Michael Dexter]

Portland Linux/Unix Group General Meeting Announcement

Who: You!
What: UnPLUG!
Where: PSU, 1930 SW 4th Ave. Room FAB 86-01 (Lower Level)
When: Thursday, March 2nd, 2017 at 7pm
Why: The pursuit of technology freedom

UnConference? UnPLUG!

I have lots of topics to discuss but you should do the choosing.

See you there.


Calagator Page: http://calagator.org/events/1250471394

Many will head to the Lucky Lab at 1945 NW Quimby St. after the meeting.

Rideshares to the Lucky Lab available

PLUG Page with information about all PLUG events: http://pdxlinux.org/
Follow PLUG on Twitter: http://twitter.com/pdxlinux

PLUG is open to everyone and does not tolerate abusive behavior on its 
mailing lists or at its meetings.

See you there!

Michael Dexter
PLUG Volunteer

___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug

Re: [PLUG] Seeking opinion on new century link 1G install

[John Jason Jordan]

On Wed, 1 Mar 2017 10:03:12 -0800
John Jason Jordan  dijo:

>On Tue, 28 Feb 2017 20:50:01 -0800
>wes  dijo:
>
>>You should remove the Netgear from the network. Once you do that, set
>>your laptop back to 192.168.0.x and everything should work again.
>
>That is not clear to me. If I remove the Netgear from the network,
>nothing will go anywhere. It has a patch cord to the new modem, and
>patch cords to a 16-port and an 8-port switch, into which are connected
>both computers, the HDHomeRun, and several printers. Looking at the
>Admin page with Firefox it lists 'connected devices' as:
>
>   192.168.1.25 Brother desktop printer
>   192.168.1.126   Laptop wireless
>   192.168.1.136   Laptop eth1
>   192.168.1.148   HDHomeRun
>   192.168.1.100   Android phone

Partial success: I succeeded in getting the Synology working by
downloading and installing the .deb file for "Synology Assistant"
provided on Synology's web site. This allowed me to reset its IP
address, so now the router sees it. I still couldn't mount it until I
remembered that I had entered it in /etc/fstab and, sure enough, the
setting in fstab still had the . . 0 . address. I changed in fstab et
voilà!

Still can't get the desktop to see the internet. :(
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug