Re: [EP-tech] ePrints server is freezing very often

Hi, John, Thank you for the reply. Your insights did help me a great deal.

I was able to resolve the Maridb issue by including the following
directive in the my.cnf file:

[mysqld]
bind-address=127.0.0.1

The above action seem to have resolved the freezing issue!

The other issue pertaining to httpd error messages, Apache2 IO write:
(104), continues to flood the error log file.

As suggested by you, I tried matching an entry from the Apache access
log file with the corresponding entry in the Apache error log file.
Here is one such entry:

207.46.13.119 - - [06/Apr/2019:19:41:23 +0530] "GET
/view/people/Kaadze=3AK=3A=3A.html HTTP/1.1" 500 4124880 "-"
"Mozilla/5.0 (compatible; bingbot/2.0;
+https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.bing.com%2Fbingbot.htm&data=01%7C01%7Ceprints-tech%40ecs.soton.ac.uk%7Cb67b61b2601b4bf1c1da08d6ba9f2a17%7C4a5378f929f44d3ebe89669d03ada9d8%7C0&sdata=6OBFDRjGj2d2XuECMvIitThibLSTtmE7rGQu2%2B5MJW8%3D&reserved=0)" - from access log file

the corresponding error log entry:

[Sat Apr 06 19:41:51.295085 2019] [perl:error] [pid 121396] [client
207.46.13.119:6010] :Apache2 IO write: (104) Connection reset by peer
at /opt/eprints3/perl_lib/EPrints/Page.pm line 78

As evident from the above entry in the access log file, the access is
coming from a robot. The access and the error log files are full of
such messages.

How do I control the robot accesses through firewall rules?

Thanks and regards, Francis




On Sat, Apr 6, 2019 at 2:42 AM John Salter <[email protected]> wrote:
>
> Hi Francis,
> I think the MariaDB errors might not be connected Ted to the EPrint exportview errors.
>
> First, the MariaDB errors - this might be useful:
> https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fserverfault.com%2Fquestions%2F393862%2Fmysql-warning-ip-address-could-not-be-resolved&amp;data=01%7C01%7Ceprints-tech%40ecs.soton.ac.uk%7Cb67b61b2601b4bf1c1da08d6ba9f2a17%7C4a5378f929f44d3ebe89669d03ada9d8%7C0&amp;sdata=WWDTSkt9Tx%2BmB6zQUA%2ByIEFdBK2G3SJZg9ywqw0xdjw%3D&amp;reserved=0
>
> Adjusting the MariaDB configuration and/or firewall setting on your server should help.
>
> For the exportview errors, this can be a result of someone (or some machine) making a request, but not waiting for a response. This can happen because the export views can take a bit of time to generate. If they take too long, a browser might timeout - and the user might try again - increasing load on the server.
>
> If possible, try and match up the entries in the Apache error log with entries in the Apache access log. You may be able to see the parameters that are being requested - and then be able to recreate the issue.
>
> It may be that the exportview request contains many EPrints, with lots of authors (e.g. some physics papers might have thousands of authors), which will take time to generate.
> If we can understand exactly what causes the slow response, we may be able to fix it!
>
> There may even be robots visiting your repository that aren't behaving nicely - which you could block with firewall rules.
>
> I hope some of that helps!
> Cheers,
> John ________________________________
> From: [email protected] <[email protected]> on behalf of Francis Jayakanth via Eprints-tech <[email protected]>
> Sent: 05 April 2019 19:09:21
> To: [email protected]
> Subject: [EP-tech] ePrints server is freezing very often
>
> Hi Everyone, In the recent past, our eprints server,
> eprints.iisc.ac.in, in becoming unresponsive very frequently - i would
> say, once every 30 minutes. Restarting the httpd server brings back
> the server, but in less than 30 minutes it freezes again! I had a look
> at the httpd error log file, and was surprised to see that the log
> file is full of entries similar to the ones indicated below:
>
> [Fri Apr 05 22:44:43.717856 2019] [:error] [pid 114613] :Apache2 IO
> write: (32) Broken pipe at /opt/eprints3/cgi/exportview line 108
> [Fri Apr 05 22:44:52.392811 2019] [:error] [pid 114910] :Apache2 IO
> write: (32) Broken pipe at /opt/eprints3/cgi/exportview line 108
> [Fri Apr 05 22:44:58.314788 2019] [:error] [pid 114919] :Apache2 IO
> write: (32) Broken pipe at /opt/eprints3/cgi/exportview line 108
>
> I also had a look at the Mariadb log file, which is bombarded with
> entries indicated below:
>
> 190405 23:23:59 [Warning] IP address '104.236.158.58' could not be
> resolved: Name or service not known
> 190405 23:24:54 [Warning] IP address '104.248.56.116' could not be
> resolved: Name or service not known
> 190405 23:25:25 [Warning] IP address '180.250.18.136' could not be
> resolved: Name or service not known
> 190405 23:26:31 [Warning] IP address '157.230.33.59' could not be
> resolved: Name or service not known
> 190405 23:26:59 [Warning] IP address '104.236.248.209' could not be
> resolved: Name or service not known
>
> All the IPs listed in the Mariadb log file have their origins in
> China, South Korea, and other places.
>
> Can someone help me to resolve the issue, please?.
>
> We are running, eprints version: 3.3.15 eps on CentOS Linux release
> 7.4.1708 (Core)
>
> Thanks and regards, Francis
> *** Options: http://mailman.ecs.soton.ac.uk/mailman/listinfo/eprints-tech
> *** Archive: https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.eprints.org%2Ftech.php%2F&amp;data=01%7C01%7Ceprints-tech%40ecs.soton.ac.uk%7Cb67b61b2601b4bf1c1da08d6ba9f2a17%7C4a5378f929f44d3ebe89669d03ada9d8%7C0&amp;sdata=A4lvW8zswjxzgxAE%2Fw1guEyMPuVREwVCDenaMAw3%2Bug%3D&amp;reserved=0
> *** EPrints community wiki: https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwiki.eprints.org%2F&amp;data=01%7C01%7Ceprints-tech%40ecs.soton.ac.uk%7Cb67b61b2601b4bf1c1da08d6ba9f2a17%7C4a5378f929f44d3ebe89669d03ada9d8%7C0&amp;sdata=fiW9E8vPtBBw%2BLawbzN5KiSkbgpLRyb65y%2BSao29r98%3D&amp;reserved=0
> *** EPrints developers Forum: https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fforum.eprints.org%2F&amp;data=01%7C01%7Ceprints-tech%40ecs.soton.ac.uk%7Cb67b61b2601b4bf1c1da08d6ba9f2a17%7C4a5378f929f44d3ebe89669d03ada9d8%7C0&amp;sdata=uUzeMlJJPkNhta1%2F5%2B07ub82MkOvs0S0XmGjmJyL2U4%3D&amp;reserved=0

*** Options: http://mailman.ecs.soton.ac.uk/mailman/listinfo/eprints-tech
*** Archive: http://www.eprints.org/tech.php/
*** EPrints community wiki: http://wiki.eprints.org/
*** EPrints developers Forum: http://forum.eprints.org/