Back to HadCM3_technical_notes
There is a BRIDGE-wide crontab that cycles through users. This needs root access. Could move to a system where all users have their own crontab.
tidy_expt does not have the same checks as ftp_master/control, and can sometimes delete files.
pi->pt and pg->pl is done during downloads, but not after tidy_expt. Webpage processing will do this final step, or it can be done manually.
search for ‘ggpjv’ in scripts as sometimes functionality only runs for user ggpjv.
You can run the downloads manually by running ~swsvalde/bin/ftp_control. This is a wrapper script that gathers arguments to pass to ~swsvalde/bin/ftp_master.
ftp_master again assembles a lot of arguments, makes appropriate directories and sshfs links, checks quotas, decides which files to download, and runs remove_unwanted_files remotely. It then loops over files to be copied. It copies the file over via the sshfs mount. It then removes the remote file. It then runs ~swsvalde/bin/convert_file with appropriate options:
${swsvalde_root}/bin/convert_file$suffix -file $file1 -expt $expt -type $type -list_runs $list_runs -check $check -famous_rename $famous_rename
$suffix can be set to is ggpjv for testing purposes.
${swsvalde_root} is ~swsvalde.
$file1 is the name of the file to be converted
$expt is the experiment name (this will have a 1 or 2 etc. at the end if this is part of a series)
$type is the type, e.g. pa, pb, pc etc.
$list_runs is y or n depending on whether the list_runs entry in ~swsvalde/ummodel/scripts/html_list/jobs exists.
$check is y or n, usually n I think in the automatic downloads, but default is y in the convert_file script I think.
$famous_rename should be y if using famous, default is “” in the convert_file script I think
For example, when doing some of the benchmarking for ib3, I ran:
~swsvalde/bin/convert_file -file xqhgba#pd000001850dc+ -expt xqhgb1 -type pd -list_runs y -check y -famous_rename n
Note that the script will move the converted file to ~swsvalde/umdata/${expt}.
precipevap folder is only used for .monthly files and for sed filss over land/ocean -> no global evap climatology field is produced.
wetland and biome post-processing id sone on oligocene in a separate queue - see in ummodel/scripts/sed2.dat.
uid: id -u or id -u [username]. ggdjl = 50892. ggdagw = 57295. swsvalde (new) = 18541731. paleo = 12825. To see the uid of a file’s owner: stat -c '%u' yourfile.txt or ls -ln
~swsvalde/scripts/make_all is the main wrapper. This calls a number of functions that are in make_all.functions for example, one of these functions is run_run_aver. This is a wrapper for the run_aver script.
${script_home}/run_aver -expt $expt -types $types, e.g. ~swsvalde/ummodel/scripts/run_aver -expt xqhgb2 -types pc,pd,pf,pg