adus
-
Posts
13 -
Joined
-
Last visited
Posts posted by adus
-
-
On 3/16/2020 at 9:45 PM, andrewatfornax said:
1. Yes. It's definitely possible to have multiple servers with the same Revive Adserver code on them, all using a central DB.
What happens with the cron job? Should it be on all the php nodes or just 1 or does it handle multiple crons going off on multiple servers?
Thanks
-
Thanks for your reply Andrew, very helpful.
-
Hi
I am setting up a fresh Revive install and I am having a brain blip. According to: https://documentation.revive-adserver.com/display/DOCS/Running+Maintenance it says:
1 * * * * /path/to/php /path/to/revive_server/scripts/maintenance/maintenance.php www.example.com
should be the cron entry, is www.example.com the url of Revive or the site I have setup for ads inside revive?
E.G. Revive install at www.blahads.com and the site its serving is www.blahsite.com which would I use?
Also if I have revive running on multiple urls such as www.blahads2.com (with 2 config files in var?) should I have a extra cron for each of those addresses or should I just have a single cron which will do maintenance for all?
The 2 urls would *run* the exact same revive install with the same www.blahsite.com inside it as well as database.I am just trying to figure out best practice etc as we use several urls to access revive.
Thanks for the help!
-
Yeah I used that page actually as a base to calculate and the eCPM was off by about 4x when I compared the numbers in revive against the manual ones I did.
-
Hi
If we sell a campaign for a weeks run the CPM calculations seem to be calculated based on a months run rather than the actual duration of the campaign.
E.g. one weeks campaign calculated a CPM of $5 but the actual number when manually calculated was $20 based on imps / 1000 then cost / that number.
Do the cpm calculations take into account the duration of the campaign?
-
They have patched it in 3.2.4 according to the git commits list. Not sure when its planned for release though.
-
We use RAS on AWS right now with the DB in RDS. No major issues although we are about 1/5 of the impressions per day but the load is minimal on that node.
Single server m3.medium runs php 8-14% cpu use. Hope that helps.
-
Yeah it looks like an update check problem. When I check manually using the web admin I get a communication failure error message.
-
I am seeing this as well on 3.2.2
-
Has this been fixed/reported to github yet?
-
Cloudflare DNS is one of the fastest out there so as long as the A record for ads.domain is with them you will get minimal dns latency. I would disable the cloudflare cache/proxy though possibly as I have noticed it lagging a bit.
What tweaks/software are you running on the server? thats your best bet for reducing perceived lag. Do you use cron based scheduling yet?
Extra redundancy
in Performance, Scalability, and Reliability
Posted
Thanks for the speedy reply!
Yeah we make heavy use of AWS and RDS so sadly cant put the code on the DB server but I always meant to double check for scaling purposes if the task could lock as most of the php nodes would just be exact mirrors.
I may spin a low end box up to handle the tasks but if I at least know it works as a mirrored then I can get under way with scaling anyway.