waldronlab/BugSigDB

Study 784 Experiments not showing but exist

Closed this issue · 43 comments

https://bugsigdb.org/Study_784 shows no experiments on the page but they do exist: https://bugsigdb.org/Study_784/Experiment_1

I tried refreshing the page under ... > Refresh but that did not resolve the issue.

This probably has the same cause as #196. We'll check this Study in particular, but a permanent solution has been elusive.

We re-saved the experiment page and it's showing up now. I'm not sure if you tried to refresh the Experiment page or only the Study page. In the future, that would be more likely to succeed.

We're going to keep researching how this happened.

@tosfos I'm not sure if this is the same bug but we're seeing something similar for https://bugsigdb.org/Study_1069 where https://bugsigdb.org/Study_1069/Experiment_3 and https://bugsigdb.org/Study_1069/Experiment_4 are not displaying. I tried refreshing both the study page and the experiment page and it's still not showing.

@cmirzayi @tosfos
Experiments 7 and 8 also have this issue
https://bugsigdb.org/Study_1079

Hello @cmirzayi
I am currently having the same issue. Experiments 4,5 and 6 are not showing but they exist.
Study link: https://bugsigdb.org/Study_940

Checking

Hello,

In this study, https://bugsigdb.org/Study_1013, experiments and signatures intermittently failed to display. Reloading the page sometimes resolves this issue, but sometimes they don't display after reloading.

Signatures for experiments 5, 6 and 7 are not displaying but exist; https://bugsigdb.org/Study_950

Signature 2 of experiment 1 in this study, https://bugsigdb.org/Study_901, is not also displaying but exist.

This study is also having the same issue;
https://bugsigdb.org/Study_915

Experiments 3 and 4 and their signatures are not showing but exist.

Signature 2 of experiment 1 here is not displaying;
https://bugsigdb.org/Study_890

Signature 1 of experiment 1, 2 and 3 are not displaying for this study;
https://bugsigdb.org/Study_1054

Signature 2 of experiment 4 is not displaying for this study;
https://bugsigdb.org/Study_1026

Experiments 2, 3 and 4 with their signatures are not displaying in this curation;
https://bugsigdb.org/Study_916

We'll check on it and see how we can resolve all of these issues permanently.

Signature 2 of experiments 1 and 3 are not displaying in this study;
https://bugsigdb.org/Study_1009

Signature 1 of experiment 6 is not displaying in this study;
https://bugsigdb.org/Study_997

Experiment 2 and its signature 1 is not displaying for this study;
https://bugsigdb.org/Study_1062

I checked all of these and they seem to be appearing now so closing this issue.

@Scholarpat is facing the same issue again. Scholastica, please report it here.

Some signatures are not displaying in this study;
https://bugsigdb.org/Study_1018

Signature 1 and 2 of experiment 2
Signature 1 of experiments 3, 4 and 5

Signature 1 of experiments 11, 12 and 13 are not displaying in this study;
https://bugsigdb.org/Study_943

We have a fix in progress. We're hoping it will be resolved by Thursday.

Signatures 1 and 2 for Experiment 13 and Signature 2 for Experiment 14 in this study is not visible
https://bugsigdb.org/Study_1086

I don’t know if this is a temporal fix or if it will work for other studies with same issues, but I noticed, when I edited an experiment with invisible signatures and saved the page the signatures appeared.

Currently all the signatures are visible for this study https://bugsigdb.org/Study_1086

I added a signature for experiment 7 but it's not showing
https://bugsigdb.org/Study_1079

Update *
So, I tried out what Anashie said she did and yes the signature displays. I'm concerned this study would have more issues like this because it is still under the category "Pages using DynamicPageList3 parser function"
Not sure if this should also display in the study

We're in a catch-22 here. None of the versions of the extension we use for dependency updates will cover all situations. But the present situation is clearly not workable. We're going to update to the latest version of the extension which will work better overall, and then we'll improve it.

The server is being restarted as part of this work. It will be down for a couple of minutes.

This issue should be basically resolved. I'll follow up with more details.

This issue is continuing. Today's export is missing nearly 1000 records. This bug has become a critical issue that is impacting our ability to work with the BugSigDB data.

Please provide more details. Is it the Signature CSV?

Signatures.csv
We are seeing that the CSV download link is returning a different number of rows each time it is run. I attached an updated CSV that has all the Signatures. We'll research a better solution.

It looks like we were running out of resources on the server. We raised the resources and it looks like that alleviates the issue. I just generated a CSV for the Signatures that downloaded the full set.

@tosfos Thank you so much. It does seem to be resolved. I was wondering if/when (hopefully) BugSigDB doubles in size over the next few years, are we going to run into this issue again?

Just wondering if there's any groundwork or processes we can put into play in the near to mid-future to prevent this from becoming a problem a few years down the line?

I think that @jwokaty's idea of tracking export sizes with an alert if thing break is a good one. See waldronlab/BugSigDBExports#34 for discussion.

Thanks Sean. I agree. Jen and I met earlier today to discuss this issue and how we can detect it faster going forward. They and I are working on implementing some detection solutions for the exports and bugsigdbr. However, since this seems to have stemmed from a resource issue upstream of the exports, I wanted to check to see if there was anything else we can do to ameliorate this once and future issue upstream as well.

@tosfos Unfortunately the issue is not resolved. I just noticed the signatures.csv is once again returning differing rows each time it's run. I ran it in the past 10 minutes and got 2081 rows, 4106 rows, and 5352 rows.

We adjusted the settings again and did a restart. It looks like the problem is resolved for now, but we need to do more research because it seems our solution is not permanent.

We believe a permanent solution is in place as of yesterday afternoon. We're still monitoring.

Have any issues been seen since 9/28?

Not that I have observed. @jwokaty you are more closely monitoring this. Anything you've noticed?

No issues yet. Let's watch it over the next two weeks.

Let's close this and create a new issue if anything happens.