Seadas 8 not responding after install

[lukecarberry]

I finally got it to work, twice. Once by clicking on the app via desktop and once via the command line. But since then, I've still had trouble getting it to not freeze on that initial "Press yes if you want to check for updates" window when I boot it up. Could it be a RAM thing? My computer only has 8GB of RAM. Could also try updating my OS. Thank you all for your help. I will keep trying, knowing that it worked twice is reassuring.

[OB.DAACx - SeanBailey]

Ah, ha! Yes, it appears to be an out of memory situation. SeaDAS is asking for 4G, but since you only have 8G there are probably times when you don't have 4G available. You can reduce the amount of RAM that SeaDAS will use by changing the -Xmx option in the <seadas-install-dir>/etc/seadas.conf file (in the line that begins with 'default_options'). Currently it is 4096M. Try 2048M or even go as low as 1024M. The lower you go, the more likely you'll run into issues when an operation needs more memory than you have allotted, but at least it *should* run

Sean

[lukecarberry]

This is good to know! I am still struggling to get it to open consistently. 98% of the time it still freezes when I open it up, even with plenty of available RAM and the -Xmx option set to 1024M. What I am trying to do is process Landsat 8 OLI imagery to get chlorophyll. Is there any reason not to use seadas 7 to do this, if it's still available? Been banging my head against this for over a week. Thank you for your help.

[OB.DAACx - SeanBailey]

SeaDAS 7.5.3 is still available and there is no reason not to use it other than it the end of the line version 7.

The processing component (OCSSW) is distinct from the visualization component and can be installed independently. So you could use the OCSSW distributed with SeaDAS 8, but visualize the outputs with SeaDAS 7. Or visualize with another tool (I'd be sad, but you needn't worry about my feelings on the matter ). You'd have to use the command line to run the processing, but it isn't really that hard to do.

Sean

[lukecarberry]

Hi Sean,

Thanks for the help. I think I have to go with SeaDAS 7 for now, but I may try to use the OCSSW for SeaDAS 8. I'm processing Landsat data for chl using l2gen. Is there any meaningful difference between the 7 and 8 OCSSWs for this task? And don't worry I'll always visualize with SeaDAS.

My other question is: Is there a guide somewhere for how to interact with SeaDAS via the command line? I am going to be processing a lot of images and would like to do them in bulk anyway. I know that there is a lot of information on how to set up the OCSSW environment, but what are the actual commands once I do this? I'll be using the Franz et al. 2015 paper as a starting point for the processing method. Thank you for your help!

Luke

[OB SeaDAS - xuanyang02]

Here is the link to the help pages of SeaDAS 8: https://seadas.gsfc.nasa.gov/help/8.0.0/

If you click -> SeaDAS TOolbox -> OCSSW -> multilevel_processor -> Command Line Usage, you 'll find the help for multilevel_processor's command line usage: https://seadas.gsfc.nasa.gov/help/8.0.0 ... or.html#cl

[gnwiii]

I am going to be processing a lot of images and would like to do them in bulk anyway. I know that there is a lot of information on how to set up the OCSSW environment, but what are the actual commands once I do this? I'll be using the Franz et al. 2015 paper as a starting point for the processing method.

The current OCSSW installer has a "--benchmark" option that installs a level-0 file and a bash script that does end-to-end processing from level-0 to a mapped image.
This is useful to check that your installation is working and also for examples of a processing workflow. The multi-level processor works well for NASA's workflows, but many workflows need other tools so I generally end up using shell scripts. If you have access to a high-end linux system you may want to investigate GNU parallel to run multiple jobs at the same time, but note that the "optimal" number of jobs is usually far less than the number of cores -- I usually start with 2 jobs and keep doubling until the thruput falls off. This can be due to I/O bottlenecks or heat causing reduced clock rates.

If you are not comfortable with shell scripts you will find the time spent learning basics of command-line processing is well spent.

[lukecarberry]

Hi xuanyang02 and gnwii,

Thank you for the help. I've been able to install the command line processing for OCSSW and now I'm trying to process a Landsat image. Again, I'm using Seadas 7.5.3 because Seadas 8 doesn't run on my computer (even though the OCSSW would technically work). I am getting the error "/Users/seadas/ocssw/src/l2gen/l2_generic.c line 458: product angstrom_1609 not found." Any thoughts? I attached a document with a full dialogue from the command line. Thank you for any help you can offer!

[gnwiii]

Which version of the OCSSW command-line tools are you using? If you have the new version, are you able to run the benchmark?

I have an older system running linux that did random freezes. Google found that other users with similar hardware were also having problems, and a bug report that identified an error in the linux kernel. I had to use and older kernel until the error was fixed in a current kernel. Some distros backport such fixes to older kernels. You may want to investigate reports from linux users of similar probelms with your hardware (graphics, network, CPU).

[OB.DAACx - SeanBailey]

I am getting the error "/Users/seadas/ocssw/src/l2gen/l2_generic.c line 458: product angstrom_1609 not found." Any thoughts?

The code only supports angstrom derived from visible and NIR wavelengths. You could simply select angstrom (as you have done) without wavelength specificity. This will provide the angstrom between 443nm and 865nm (or the closest bands to those for the instrument).

Sean

P.S. Since this is a separate topic from the original issue, you should post a new question.