fix disk mounting and other startup sequence commands
added the deadman scripts to install
Updates to quiet warnings and errors on Perl 5.30. No functional change.
[amztpp] Try to fix the problem that if a user edits a comet.params (or other type of ) file during a session, the changes would not be re-uploaded to S3. I tried fixing this using the current putList() framework, but was thwarted. In the end, I just removed a section for intercepting files that had already been uploaded and just always check. This may be slower. But perhaps safer. And I couldn't get it to work with putList(). ParallelManager seems to stomp on putList() is a way I don't under...
[amztpp] Just added a comment and floundering around for a long time
remove debugging print
back out a hack that I introduced to get instances to start faster. would likely cause problems. Need a better way to solve this problem.
add a new credentials based instance tag and set getInstances to only return instances for the current credentials. getAllInstances is now available to list all instances even those belonging to others on different credentials.
add a new 'status' queue. It doesn't do anything yet.
tweak the text based format to show more information about S3 contents and instances running
Add option 's3clean' to just remove files in this S3 bucket
Better instance status table
Now specify the region and location for bucket creation.
yes, we DO need to specify the region after all here
remove two obsolete/removed files from the MANIFEST
It is not clear why messages that are already handled can still sit in the queue. But they do. So, for now, after a message is found to fester in the queue for 30 polls, go ahead and try to handle it again. Unclear if this is the right thing to do or not. SQS is temperamental.
Since AWS no longer charges in increments of full hours, make the default to shut down after 3 minutes with nothing to do, rather than wait
It seems to take quite a while before the manager sees that an instance is up and working. Therefore waiting to start new instances until the previous one is known to be searching leads to a long ramp up. Allow the system to start up to 5 new EC2 nodes ahead of where they are known to be up and searching. This is not an ideal fix. Need a better way to compensate for such slow feedback.
Fix handling of mzML.gz files that remain compressed during processing.
Since AWS no longer charges in increments of full hours, make the default to shut down after 3 minutes with nothing to do, rather than wait around until the uptime is close to a full hour increment.
Don't uncompress mzML.gz files since we search them compressed with Comet and they are usually compressed on our end already. It would be a bit better to know if they were uncompressed at the source and then uncompress them, but if compressed at the source, then leave them compressed. That would like require adding to the passed message.
Nicer wall message for auto shutdown
Remove deprecated UpStart system init scripts, and add replacement systemd init scripts for managing amztppd and deadman services
- Added hack around weird things in the internals of Amazon::SQS::Simple
- Widen display for modern ids
- Comment out some hacks to try to get S3 working better. They don't obviate the Net/Amazon/S3/HTTPRequest.pm hack as I thought they should. But maybe still along the right path?
- Fix group_description to group-description
Copied newer and apparently better code from Joe's branch amztpp-2.4.4, which appeared to be already semi-active on regis. Yet to be tested.
Fixed copy paste error
New script to replace the original deadman by jslagel. Wait a certain period of time and then shut down the system. Other active processes should kill and reset this script while it is waiting to indicate that they are still active.
Hi John, no need to migrate PSI-MOD to GitHub, it is already there! https://github.com/HUPO-PSI/psi-mod-CV Can you help with an outstanding issue: https://github.com/HUPO-PSI/psi-mod-CV/issues/5 Do you have the script that generates the XML from the OBO? If so, would you share it? If you send me your GitHub username, I can add you to the organization. Thanks! Eric From: John S. Garavelli jsgaravelli@users.sourceforge.net Sent: Monday, February 10, 2020 10:16 PM To: [psidev:mod-controlled-vocab-changes]...
Hi Val, we are now maintaining PSI-MOD over on GitHub, This site is now obsolete. I have posted your issue here: https://github.com/HUPO-PSI/psi-mod-CV/issues/9 I hope that it will be addressed very soon. Please continue the discussion there and create new term requests as GitHub issues over there. thanks! Eric
Added Gerben's suggestions
added original submission zip archive for refer...
Added todos for the example file
Added a TODO text file
accepted tracked changed and clarified insertion.
removed non-OBO format comments at bottom and t...
added examine from original submission. this ha...