| Age | Commit message (Collapse) | Author |
|---|
|
|
|
There is a problem with long paths on windows, which causes some of
the ttb logs in this suite not to be created. To go around this, the
original priv_dir from the Config is no longer used for writing the
logs. Instead a new priv_dir is created in the data_dir - which makes
the path much shorter.
There is also a problem caused by the lower resolution of the system
clock on windows. It makes the test cases for sorting trace messages
fail. To get around this a sleep of 2 ms is added in "appropriate
places", and also the messages sent between client and server when
creating the trace log for these test cases is now better synched.
The cleanup functions, which terminate slave nodes, was called in
end_per_testcase. However, it seems to be a bug in the test_server
which causes this to hang if the test case failed with a
timetrap_timeout. Workaround for this is to do the cleanup in
init_per_testcase instead - i.e. make sure that nodes that are to be
started by the test case do not already live when the test case
starts.
|
|
Slave nodes were earlier stopped inside each test case. If a test case
failed before this point, a slave node would survive and it might
interfere with the next test case causing multiple failures. This
commit moves the stopping of slave nodes out to a separate function
for each test case - called during end_per_testcase.
A minor correction is also done in ttb:ensure_no_overloaded_nodes -
the reply message sent back from the ttb process is tagged so only the
expected message will be picked from the message queue. Otherwise, for
instance nodedown messages from the monitoring of slave nodes (by the
test cases) could be received here.
Finally, the sleep timer when waiting for trace messages to arrive
over tcp/ip is extended a bit since test cases sometimes failed with
missing trace messages here.
|
|
OTP-9430
OTP-9403
OTP-9431
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The r11 option is no longer supported by the compiler (silently
ignored).
|
|
|
|
|
|
Also, remove compiler warnings for crashdump_viewer_SUITE and fix
Makefile in test directory so Emakefile does not grow.
|
|
This is a first attempt at fixing the problem described in seq11783 -
crashdump_viewer is very slow at parsing big crashdumps. To open the
first page for a dump of 17M takes about 2 minutes and a dump of 280M
takes 1.5-2 hours.
The main problmem is that the cdv_dump_index_table, which holds all
tags read from the dump, is a bag. Profiling shows that ~95% of the
time is spent in ets:insert. The table is now changed to an
ordered_set.
A second problem occured when a page with many table rows was
opened. These pages were sent to inet in one chunk, causing both
crashdump_viewer_server and the inets (mod_esi) process to grow very
much in memory usage. To overcome this, the pages are now sent to
inets in chunks of 1000 rows, and the data is coverted to binaries to
avoid data copying between the two processes.
Also, some new information in the crashdump was not recognized by the
crashdump_viewer. This has been fixed.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
