]
Roland Grunberg updated JBDS-4085:
----------------------------------
Attachment: disable-uses-by-default.patch
unresolved bundles, slow initial start up
-----------------------------------------
Key: JBDS-4085
URL:
https://issues.jboss.org/browse/JBDS-4085
Project: Red Hat JBoss Developer Studio (devstudio)
Issue Type: Bug
Components: rpm
Reporter: Václav Kadlčík
Assignee: Nick Boldt
Fix For: 10.2.0.AM3
Attachments: disable-uses-by-default.patch, disable-uses-by-default.patch,
restart-after-installing-m2e-wtp-jsf-stuff-from-Central.txt,
restart-after-installing-m2e-wtp-jsf-stuff-from-Central_full.txt,
rh-eclipse46-devstudio10.2.log.20161028-1611.txt, unresolved-max.out, unresolved-mid.out,
unresolved-min.out, workspace-max.log, workspace-mid.log, workspace-min.log
Tested:
* rh-eclipse46-devstudio-10.2-0.20161004.2205.el7.x86_64.rpm
* rh-eclipse46 Beta
(
http://download-node-02.eng.bos.redhat.com/rel-eng/RHSCL-2.3-RHEL-7-Beta-...)
rh-eclipse46 has its own unresolved bundles
(
https://bugzilla.redhat.com/show_bug.cgi?id=1381820). But when you add
rh-eclipse46-devstudio, the number of unresolved symbols increases quite dramatically and
initial start up times also get worse. (I guess that the two problems can be connected so
I file them together but feel free to split them if I'm mistaken).
I'm attaching "ss -s INSTALLED" and workspace logs for several scenarios:
* "min" in the name means minimal install rh-eclipse46-devstudio
* "mid" is for "yum install rh-eclipse46 rh-eclipse46-devstudio"
* "max" is for "yum install 'rh-eclipse46*'" - all the
available RPMs, including devstudio
As for start up times, it looks like devstudio adds ~80s for one particular machine (Xeon
with 16 cores 2.40GHz, 16 GB RAM). For example:
initial start of rh-eclipse46: 40s
same as above with rh-eclipse46-devstudio: 120s
initial start of rh-eclipse46* except devstudio: 50s
same as above with rh-eclipse46-devstudio: 140s
Roland Grunberg managed to improve initial start up times in rh-eclise46 a few weeks ago
(
https://bugzilla.redhat.com/show_bug.cgi?id=1362178#c17). I don't know it it helps
here but I'm linking it anyway...