Easyfind down?

[Ubi]

For some reason the easyfind script on the server does a hardcoded remote lookup @54.247.92.107. Afaik this IP is not owned by Excito (although could be, it's amazon web services). The lookup seems to be related to JSON validation, but the serverside code has no documentation whatsoever so Excito needs to comment on this.

Host 54.247.92.107 is currently not responding on port 443 (although the host itself is up), which means that Easyfind is effectively disabled for everybody until either the host comes back up or the code is changed to point to a different IP.

In either case, having this secondary requirement on the easyfind service using a hardcoded IP does exponentially increase the potential for downtime, so this may need to be addressed at a more thorough level.

[Gordon]

Seems like a backoffice server is down again. Only this time it spits raw data from the error that this invokes on the update script, which is probably not as intended and the result of debugging the previous issue, just over a week ago.

I suggest you edit that last post and cross out those two parameters from the validate_production() function call. Although they're not fully complete, they contain a lot of information already which you should not share.

[wm.bubba]

Thanks for the tip Gordon. Post has now been edited.

[Ubi]

Seems like a backoffice server is down again. Only this time it spits raw data from the error that this invokes on the update script, which is probably not as intended and the result of debugging the previous issue, just over a week ago.

no, the hardcoded IP was in there since at least september 2013. Although the original problem was related to this same piece of code, the current problem is new and unrelated. Last time the backoffice server was not down, now it is.

[johannes]

Ubi is right, its a secondary server that is not responding now, handling the validation of the b3 requesting the easyfind update. The reason seems to be a failed security update, we'll investigate immediately tomorrow morning.

Note though that Easyfind is working, again only IP updates are failing.

Sorry about the downtime again.

[Gordon]

Obviously I don't own a map of the whole Excito infrastructure. The previous error appeared to be a json version of a HTTP 500 error, which can mean just about everything but usually indicates a timeout on a secondary service if the same procedure normally functions okay.

This last error is different in the sense that it barfs a lot of debug information, showing exactly what is wrong and I asked myself out loud whether the update script was actually supposed to do that or return a generic error like before.

[johannes]

Yes, Gordon you are also right, the last time the error was a full disk (log file massacre), this time it appears Apache didn't restart after a security update we did last thursday, which we failed to remedy immediately. Should have been a simple fix but as you might know Excito is kind of short of people, and the persons designing this system are no longer around, making these things a bit tricky. I hope we can fix it during tomorrow, I have gotten some pointers from Carl (who built the system) to help us out.

I am not very proud of the Bubba IT system's general shape now, but there are no resources for an (obviously much needed) thorough check of everything. We'll have to patch it up as good as we can.

Thanks for understanding and as always, your help and presence here on the forum is invaluable.

[johannes]

Should be up now!

[wm.bubba]

I can confirm it's working now

Once again, thanks Johannes.

[OlivierC]

Instead of using Web GUI, i have used script and i can confirm it's working.

Code: Select all

/usr/lib/web-admin/easyfind.pl

Output

Updating IP on file.
Wrote config to file
{"opcode":1,"record":{"domain_id":"1","content":"my_previous_ip_here","name":"my_B3_name_here.myownb3.com","type":"A","id":"2288"},"error":false,"operation":"IPupdate"}