- Tuesday June 22 2010, 08:27 -
- rama has joined the room
rama: about the refresh.pl sccript
sjcarbon: ah, yeah.
sjcarbon: ben: from that bug it looks like the refresh scripts aren't running very frequently (the ids have gone out of sync).
sjcarbon: Is refresh.pl run periodically? And if so, what schedule?
- 08:33 -
- ben has disconnected (Read error: Operation timed out)
sjcarbon: He escaped!
rama: no Ben yet
rama: no body else on the chat
sjcarbon: Ah well.
rama: about testing
sjcarbon: I was going to see what people's schedule looked like for the next couple of weeks.
rama: i am going on vacation on june 24th
rama: i am out for 3 weeks.
sjcarbon: Have fun!
rama: I will have email access and will be able to test if there are no firewall issues/connectivity issues
sjcarbon: I'd hate to have you test while you're trying to relax.
sjcarbon: We can see who else will be around.
rama: (relax in India?? with inlaws and family??
rama: i need a break from family. so escaping with work as excuse is a good thing
sjcarbon: Okay with me then.
sjcarbon: I hope it's not too hot where you're going.
- 08:39 -
sjcarbon: I'm hoping to scoop up the last two bugs while we're testing.
rama: it is hot there 35C
rama: that will force me to stay home
sjcarbon: I can understand the need for internal escape then.
sjcarbon: There were some caching issues that I want to talk to Ben about, but I can get him on email.
rama: i will sign off
sjcarbon: And follow up with the refresh.pl calling.
rama: thanks a lot
sjcarbon: Cheers--I'll talk to you later.
rama: one more question
- Zenbitz has joined the room
rama: how is the new PANTHEr family display going to work with the existing Homolog set displays
- Zenbitz is now known as ben
rama: welcome back Ben
ben: uh... what did I miss...
rama: nothing much
ben: should I change the cache script command invocation/timing?
sjcarbon: Rama: I don't know how the displays will integrate.
sjcarbon: Ben: What's it at now?
ben: 00 12 * * 1,3,5 /share/goweb/www-data/html/dev/amigo/refresh.pl -c
sjcarbon: Rama: I thiought the news ones were to replace the old ones?
ben: that was the last thing I chatted... I think I got discon'd when I turned on vpn to check crontab on goweb1 from home!
- gwg has joined the room
- 08:45 -
sjcarbon: Ben: Hmmm...I'll have to check the version that you're running, but I think that you'll need to get rid of the "-c" to regenerate the RG stuff with a new database.
ben: I meant to change this when we went to weekly (since the wed/fri runs don't do anything... doesn't the RG stuff take a long time?)
rama: Seth- about ref.genome data
rama: the new thing can't replace the old ones i think
rama: data needs to be merged somehow
sjcarbon: Ben: I think on the order of five or six hours--but it could be speeded up significantly (for you and the user) by redoing the cache.
sjcarbon: And to connect this to Rama's point...
rama: i am clueless about how all the ref.genome data are going to come together in amigo
sjcarbon: I didn't do that because I thought that Sven's stuff was a replacement.
sjcarbon: We can cover that with Chris at the next AHWG.
rama: need to talk to Pascale/Kara and Chris
sjcarbon: Or next RG meeting.
rama: sure. i will let pascale know
sjcarbon: Next RG then? Okay, thanks.
- 08:50 -
sjcarbon: The cache redo for RG would take a little elbow grease, so let's not worry about it right now.
rama: not worry about ? (sorry i did not follow)
sjcarbon: Rewriting the caching mechanism for the RG stuff (which would speed up refresh.pl).
ben: well I will wait until you send me the command
ben: I think we discussed priming it priot to load? or by flipping servers?
sjcarbon: I think that was in reference to the Lucene stuff.
ben: yeah, you are right
sjcarbon: The RG regenerations can be done on a live machine.
sjcarbon: Could anyways.
rama: (okay got it)
rama: i have to go now.
sjcarbon: see ya!
rama: Seth- did you have any questions for me?
sjcarbon: no,. I'm good--have fun!
- 08:56 -
sjcarbon: Ben: the RG data will fall out of sync unless the script is run without the "-c" (or wit hthe addition of "-s -p -g").
ben: we good?
ben: we need to run -c after every load
ben: should we just run "no opts" (which is everything I recall) after loading?
sjcarbon: that would be great.
ben: I could make it earlier on monday (load usually finishes ca. 6 am, so I could make it 9)
ben: I will run it right now manually to check timing.
sjcarbon: okay,great--let me know how long it takes.
ben: running... brb
sjcarbon: if it's insane on your machine, maybe i'll try and redo the caching right now.
ben: well if it takes more than 30 min, then I won't be able to report back...
sjcarbon: i'll no doubt be talking to you later then ;-)
sjcarbon: definitely slower than that.
sjcarbon: I think that covers everything that we can here...
sjcarbon: anybody else with anything?
- 09:01 -
sjcarbon: all good then--thanks everybody!
ben: Oh, sure is.
ben: I better put it in background.
ben: it might bomb anyway when I kill my terminal
ben: I think linux is more forgiving of leaving off the nohup though
ben: or I could leave my lap top here and let it go
ben: prolly do that
sjcarbon: You can rerun under screen too--that's usually what I do for things like this.
sjcarbon: Nicer than nohup.
ben: yeah but I can never remember all the screen commands
sjcarbon: byobu is a great wrapper if you have it.
sjcarbon: much simpler and with menus if necessary (it actually what I use).
- 09:07 -
sjcarbon: i'm good here, anybody with anything else?
- rama has disconnected (Quit: Computer goes to sleep!)
sjcarbon: okay--thanks all!
- gwg has disconnected (Read error: EOF from client)