Pavlovia is currently offline while the final merge goes ahead. The final merge of data from pre/post fire is currently under way. We’ll be back online as soon as possible - should be up and running this afternoon.
Update: Pavlovia back online and ready for use
Wednesday March 24?
HI All,
I started collecting data a couple weeks before the fire, and since the fire all my experiments are gone from the “Dashboard” on the “regular” Pavlovia system. I looked on tmplovia, and they’re all there on that server. Do I assume that they won’t appear on the regular pavlovia until after this next merge because they weren’t up there before Jan 3? Just wanted to be sure I’m understanding correctly.
Thanks.
–JF
Will I be able to access my account after this final restart? (I created my account after Jan 4 and have not been able to access it since the fire).
My data is still missing. My experiment is on my dashboard but is showing no results, when I had results before.
Your experiments/accounts will most possibly be restored tomorrow. They postponed today’s “restart” for some reason. Today there will be downtime to facilitate the merge process, however the real merge will happen tomorrow, it seems.
Yes, assuming we have no further hiccups it will be on pavlovia tomorrow afternoon. If you need to get access to the accounts/data before tomorrow then you can do so by logging in to gitlab.tmplovia.org instead - that has the accounts/data from Jan/Feb on it (and that now has most of the post-fire data on it as well).
There wasn’t any downtime today - the downtime is just needed for the final sweep of merging. Currently, because people are continuing to collect data continuously (which is fine) the merge never stops. So the bulk of the data have already been merged from post-fire (= pavlovia) to what is, currently, tmplovia and that will continue during today. Then tomorrow morning we need to shut down access while we perform the final merge to make sure that all changed files have been copied over (and with nothing changing during that period - that’s the need for shutting down access).
Hope that makes sense now!
Sorry, I now realise there was a typo in my post, picked up by @SinclairAmend which is that it’s Thurs 25th March that the switch will occur. I had written Weds 25th March which doesn’t exist!
I have tried accessing my account (created aft 4th Jan) on the temporary server and keep getting a password error - I don’t know whether it’s real (meaning I genuinely have the wrong password) or I am not able to access for some other reason. Can you let me know how to resolve this? Thanks!
Hi @delicato I can see your account there and about 16 projects created 2-4 weeks ago so my guess is indeed that this is a simple password problem. I’ll get in touch off-list
Thanks for the update! Do we have any idea how long we should expect Pavlovia to be down? Sorry if this has been answered elsewhere, I looked but didn’t see anyone ask this yet. We avoided scheduling participants for the 24th per the previuos post but we do have families with children scheduled for the 25th - if we cancel on them a second time in a row they’re going to start thinking we’re giving them the runaround here…
Sorry that we had a bit more delay; we encountered a couple of merging issues in our tests. We wanted to be sure that’s all ironed out before applying it for real. We expect it’s done tomorrow, but to play it safe you could add another day (so that in case we encounter another unexpected day, it won’t force you to delay again). In that case it would be Saturday.
Sorry, just to clarify, is the hope/expectation for this to take something like, 5 hours starting from 7am UK time, or more like 24+ hours from 7am UK time? If the former then there is a chance we wouldn’t have to reschedule anyone. The first cancellation (from the fire) already cost us 50% of our previously-scheduled participants who decided they weren’t interested in rescheduling. (And we already spent quite a lot of time coming up with our alternate solution to ensure the experiment could be run while the previous server was down, which we spent time on specifically to avoid this problem! 
)
Hello @HEKragness ,
I am very sorry to hear that your participants were unwilling to reconvene.
We are expecting the final stage of the recovery process to take roughly half a day. Fingers crossed!
Alain
Thanks! Definitely helpful to have a point of reference!
Thank you and the team for all the hard work you’re doing. I think it’s incredible that you’ve managed to fix it really quite quickly in the grand scheme of things, and we should remind ourselves that we’re really lucky that all the data was still there. Plus, although there are slight delays with merging, online data collection takes just a fraction of the time it does with face-to-face data collection.
Patience is key, and you guys are doing a wonderful job, so thank you.
Hi, It’s online now. So everything is good to go? Thanks for working on this!
We are back up indeed, currently testing that all is well.
We will post a new message to give additional information, in just a few minutes.
Please use caution: we had to merge over 4000 experiments and so, inevitably, there might have been some awkwardness. It is also possible that the status of your experiment need to be changed to piloting or running as it might have switched to inactive during the merge.
I’d like to add that regardless of the status of the merge, we have a full copy of the pre-fire data and a full copy of the post-fire data before the merge. So we can work with you to restore your projects to an adequate state.
