Now make_file_index()
takes optional argument index
, which
should specify the index of a previous version of the input
file. If specified, it will only index the end of the file that is
not already indexed. This provides a quick way to update the index
of a file that is continously growing.
Add sge_make_date_index()
.
Add sql_import_sge_accounting()
.
Add beegfs_quota()
.
Now read_raw_sge_accounting()
and ead_sge_accounting()
can read
the subset of data that is given by an
sge_accounting_index_by_week
week index table. This is one by
passing to to argument offset
.
Now read_raw_sge_accounting()
and read_sge_accounting()
accepts
also file connections as input.
Add sge_failed_codes()
.
Updated SGE accounting vignette.
Updated example data for number of users over time.
disk_free()
failed because some columns were not recognized.
Field nbr_of_jobs
returned by sge_make_week_index()
was a
cumulative sum, not number of jobs per week.
open_file_at(..., auto_close = TRUE)
was not safe; it could end
up closing the incorrect connection if the intended one had been
closed and another one had been opened. Dropped options for
auto_close
to be on safe side.
Add sge_make_week_index()
.
Add add_weeks()
.
Add users_over_time()
.
Add generic function anonymize()
with implementations for
data.frame
, raw_sge_accounting
, and qstat
objects.
Add read_qstat()
for getting information on the jobs currently on
the SGE queue.
Add make_file_index()
, save_file_index()
, and
qread_file_index()
.
Add open_file_at()
.
read_raw_sge_accounting()
, and therefore also
read_sge_accounting()
, gained argument offset
.
ru_wallclock
values returned by read_sge_account()
were
corrupted by "random" NAs.Add write_raw_sge_accounting()
.
Add anonymize_raw_sge_accounting()
.
Now read_raw_sge_accounting()
records file header in attribute
header
.
Added an example SGE accounting
file and examples how to process
it.
job_summary_by_status()
.parse_category()
for sge_accounting
objects.read_sge_accounting()
to display how to inspect
how much CPU time was spent on successful and failed jobs.R CMD check --as-cran
on systems without qhost
.Add sge_accounting_file()
and read_sge_accounting()
.
Renamed host_table()
column # Physical Cores
to Physical
Cores
so that it does not conflict with #
-comment lines in
tab-delimited files.
The system("scripts", "system_info.R", package="wyntonquery")
script is now robust against errors; if there's an error querying a
particular host, then that host is dropped from the final out.
The system("scripts", "system_info.R", package="wyntonquery")
script now caches results from individual compute nodes.
host_info()
now reports on cpu_model
, which corresponds to the
cpu_info()
field model_name
, e.g. "Intel(R) Core(TM) i7-8650U
CPU @ 1.90GHz". Previously, it tried to report on the CPU speed
but that fluctuated from run to run because it was incorrectly
inferred from the cpu_mhz
field.
on_hostname()
gained argument cache
for controlling whether
cached results should be considered or not.
It's now possible to specify which hosts to query via environment
variable R_WYNTONQUERY_INCLUDE
.
on_hostname()
gained argument on_error
for controlling how
errors should be handled.
cpu_info()
did not handle upper-case letters in field names.Add host_table()
.
Add round_ram()
.
Add lscpu()
.
Add system_info()
.
Add disk_free()
.
Add queues()
.
Add on_hostname()
.
Add qhost()
and cpu_info()
.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.