You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've messaged @sizzlemctwizzle ... I rolled back the commit HEAD on production and it's still only showing minimal pages... he'll have to check out S3 storage which I don't have access to.
Arggh... so assigning to @sizzlemctwizzle ... I don't know how to fix this... sizzle mentioned that we might drop S3 and move it to production to avoid intermittent S3 issues.
sizzle said that my Hello, World! Unit Test script is present on S3 (I republish that usually after every dep update) sooooooooooo... that probably means that we still have the source at least just no linkage to it yet... at least I hope the source is all there. crosses things
It just makes me lose so much faith in mongoose. I mean, who in the hell
can get it that wrong? A method that is supposed to remove one document
from a collection removes ALL of the documents in that collection. It's
just ridiculous.
Well a month ago is better than starting from scratch I think. I'd like to
keep the published dates intact since that is in the _id currently for
scripts. We'll lose the users added since then as well... they can always
resign up though... I'm feeling awful right now... I've added remove (delete) ops to my normal test routines as well.
We'll have to keep this blocking until I remigrate #285 #285 on that DB ...
and checkout an older HEAD here.
—
Reply to this email directly or view it on GitHub #729 (comment)
.
Would it be useful to post a list of scripts that don't have any metadata page, which makes them difficult to find? I haven't noticed that the site exposes any numeric index for simple spidering.
Contributing is always welcome... I do have my own opinions on the direction as well as being the Active Maintainer... a lot of my thoughts on direction just changed with this issue too... some didn't.
One thing though is we don't use SQL... it's mongoose or MongoDB.
And don't forget to test them once in a while too.
Would it be useful to post a list of scripts that don't have any metadata page, which makes them difficult to find?
I presume these exist and those authors need to come forward and let me know what to look for. I can probably, eventually, be able to look it up... but I need their @name and their username to do this... without this they are orphaned on S3... this is what sizzle means by the "DB is devastated"... we lost all "linkage" due to that dependency deleting the entire collection instead of a single document (record). As I pointed out to @sizzlemctwizzle this could have been horrifically worse had we had a "delete user" option, as suggested in #359, and someone deleted themselves and poof all the users would have gone bye bye including myself. :\
I haven't noticed that the site exposes any numeric index for simple spidering.
SEO is handled differently than GreasyFork (GF) and we use username/scriptname in the route. I'm not real fond of looking at GF links as most of the time they are completely useless by not telling you anything until you visit that site. I'm open to improving SEO rankings but that would be another issue (and milestone at SEO) . I seriously doubt a plus vote will happen from me on changing this gently suggested routing url.
@sizzlemctwizzle I still can't launch local pro possibly due to what you said at #712 (comment). Since the DB in use is 21ish days old it is pre this comment. I need this access to migrate the DB over the hump to get to 0.2.0 of our project and any manual look ups for those authors that may need their code from S3 slipped in somewhere (probably not the site itself)... so far no one has made any requests yet.
This user for example registered in the last 21ish days but didn't make any requests to be restored script source wise.
@Martii
I'm not that interested in taking a role, per se, at least not at this point. If SQL is ever required, desired, or seems like it might be nice, i'm willing to help this excellent site.
We'll be here when you are ready. It's okay to take part in issues as well... it is helpful to get a different perspective on what everyone wants versus what can be delivered. OUJS is, in part, about safer choices instead of closed source and/or forcing unreasonable .user.js engine features out of the way like some sites do.
If SQL is ever required
I'm open to examining this at the node level myself... mongoose is a different type of DB than I'm normally used to (I use "squeekel" myself for some applications ;)... if the benefits outweigh the effort, OUJS as a whole will surely consider it.
Activity
[-]Number of scripts 0[/-][+]Number of scripts: 0[/+]Martii commentedon Sep 1, 2015
Give me a few to investigate this.
Site refs:
https://openuserjs.org/discuss/All_scripts_deleted(lost in restoration)https://openuserjs.org/discuss/Search_doesnt_work,_only_scripts_posted_in_the_last_few_hours_are_listed_in_main_page,_and_even_clicking_groups_are_empty(lost in restoration)Martii commentedon Sep 1, 2015
I've messaged @sizzlemctwizzle ... I rolled back the commit HEAD on production and it's still only showing minimal pages... he'll have to check out S3 storage which I don't have access to.
Martii commentedon Sep 1, 2015
@sizzlemctwizzle
dev currently still has it's full compliment ...
at everyone else... still awaiting to hear back from him... this might take some time if he's not available. Sorry.
Downgrade the two of the three latest deps
Martii commentedon Sep 1, 2015
@sizzlemctwizzle
Nada... downgrading the (affected) deps still isn't showing full compliment on local pro and actual production.
Martii commentedon Sep 1, 2015
Arggh... so assigning to @sizzlemctwizzle ... I don't know how to fix this... sizzle mentioned that we might drop S3 and move it to production to avoid intermittent S3 issues.
Martii commentedon Sep 1, 2015
Added blocking since this is definitely blocking... may modify CONTRIBUTING.md to allow the Active Maintainer to add for ESR type situations.
Martii commentedon Sep 2, 2015
OH NO!!!!!! :(
@sizzlemctwizzle
Ref: Automattic/mongoose#3326
Merge pull request #727 from Martii/someDepUpdates
Martii commentedon Sep 2, 2015
sizzle said that my Hello, World! Unit Test script is present on S3 (I republish that usually after every dep update) sooooooooooo... that probably means that we still have the source at least just no linkage to it yet... at least I hope the source is all there. crosses things
2 remaining items
sizzlemctwizzle commentedon Sep 2, 2015
It just makes me lose so much faith in mongoose. I mean, who in the hell
can get it that wrong? A method that is supposed to remove one document
from a collection removes ALL of the documents in that collection. It's
just ridiculous.
On Wed, Sep 2, 2015 at 12:07 AM, Marti Martz notifications@github.com
wrote:
chacham2 commentedon Sep 2, 2015
May i help with any SQL related issues?
And don't forget to test them once in a while too. Those are 2 mistakes people don;t realize until after they needed them. :)
BTW, oujs is awesome. I hope you get this all cleaned up!
jscher2000 commentedon Sep 2, 2015
Would it be useful to post a list of scripts that don't have any metadata page, which makes them difficult to find? I haven't noticed that the site exposes any numeric index for simple spidering.
Martii commentedon Sep 2, 2015
@chacham2
Contributing is always welcome... I do have my own opinions on the direction as well as being the Active Maintainer... a lot of my thoughts on direction just changed with this issue too... some didn't.
One thing though is we don't use SQL... it's mongoose or MongoDB.
That is sound advice. Thank you.
@jscher2000
I presume these exist and those authors need to come forward and let me know what to look for. I can probably, eventually, be able to look it up... but I need their
@nameand their username to do this... without this they are orphaned on S3... this is what sizzle means by the "DB is devastated"... we lost all "linkage" due to that dependency deleting the entire collection instead of a single document (record). As I pointed out to @sizzlemctwizzle this could have been horrifically worse had we had a "delete user" option, as suggested in #359, and someone deleted themselves and poof all the users would have gone bye bye including myself. :\SEO is handled differently than GreasyFork (GF) and we use username/scriptname in the route. I'm not real fond of looking at GF links as most of the time they are completely useless by not telling you anything until you visit that site. I'm open to improving SEO rankings but that would be another issue (and milestone at SEO) . I seriously doubt a plus vote will happen from me on changing this gently suggested routing url.
@sizzlemctwizzle
I still can't launch local pro possibly due to what you said at #712 (comment). Since the DB in use is 21ish days old it is pre this comment. I need this access to migrate the DB over the hump to get to 0.2.0 of our project and any manual look ups for those authors that may need their code from S3 slipped in somewhere (probably not the site itself)... so far no one has made any requests yet.This user for example registered in the last 21ish days but didn't make any requests to be restored script source wise.
chacham2 commentedon Sep 3, 2015
@Martii
I'm not that interested in taking a role, per se, at least not at this point. If SQL is ever required, desired, or seems like it might be nice, i'm willing to help this excellent site.
Martii commentedon Sep 3, 2015
We'll be here when you are ready. It's okay to take part in issues as well... it is helpful to get a different perspective on what everyone wants versus what can be delivered. OUJS is, in part, about safer choices instead of closed source and/or forcing unreasonable .user.js engine features out of the way like some sites do.
I'm open to examining this at the node level myself... mongoose is a different type of DB than I'm normally used to (I use "squeekel" myself for some applications ;)... if the benefits outweigh the effort, OUJS as a whole will surely consider it.
Martii commentedon Sep 4, 2015
@sizzlemctwizzle
Please confirm we have a working back up of the re-migration later this weekend, unblock this issue, and close at your leisure. TIA
sizzlemctwizzle commentedon Sep 4, 2015
Will do.
On Sep 4, 2015 12:37 AM, "Marti Martz" notifications@github.com wrote:
sizzlemctwizzle commentedon Oct 3, 2015
Confirmed that we have one backup (created at midnight UTC) per day for the last 7 days. It is fairly easy to restore from one of these backups.
Martii commentedon Oct 3, 2015
Gratsi! :)