Open R0n1n-77 opened 8 years ago
In addition, browsing to the hash url of the corresponding request like:
/turpentine/esi/getBlock/method/esi/access/public/ttl/86400/hmac/f19dbb753bd2f245f19a562e9beebd6d3211cf5736da12519b596be44eca3469/data/qoB3aU7bq-oZ3EIZMePYpOvkRXoLLY.nu-Y.2Vi2uUddhFoSh789sDCdy1zicSnfZ1KT8SyHfcapVpnse0OfjHZ7PSgUSEs8oMbLROCGiAXMzbNLU2v8EP1
Returns - "ESI data is not valid". Other logs like, PHP, Apache and Magento do not produce any errors. Please advise.
~~Noticed another thing. If I remove the real server IP from a "Crawler IP Addresses" and leave just 127.0.0.1, then crawler starts working as expected. But then warmer script becomes useless.~~
Actually found out, that crawler starts throwing errors when the warmer script is running. Is there any config that may cause them to conflict?
I haven't experienced this one before. It sounds like an issue occurring when Varnish attempts to connect to the back-end. This seems to make it sound like it might be related to having gzip compression set up for nginx.
@aricwatson I don't have gzip enabled in Nginx, although traffic to backend is really proxied through it to apache. Gzip in apache is enabled. Adding a document to .request, as suggested in a link you've sent looks like made a change. Crawler is still throwing errors on more then half of the URLs, but URLs are actually being crawled, F>or I'm getting "X-Varnish-Hits 1".
Crawler is still throwing errors on more then half of the URLs,
And it's still the broken pipe error?
Yes, right. Varnish still returns the broken pipe error.
Hi, I noticed that crawler stops working randomly. Flushing Magento cache and restarting Varnish fixes it, but issues starts occurring again after half an hour or hour. In a turpentine log I get:
In a Varnish log:
Please advise.