Last modified: 2012-03-29 16:46:06 UTC
A running TestSwarm browser client seems to fairly routinely encounter unrecoverable errors when reloading its page. Most common I've seen have been: "No input file found" "Bad gateway" from the web server / cache frontend After a weekend for instance, usually I'll have several browsers down; sometimes it'll happen a couple times in a day too. Since the result is that none of the web page or scripts get loaded, it's unable to automatically retry -- someone has to come along and manually hit 'reload' on the browser. If instead of reloading the entire page, it reloaded an iframe or loaded data over ajax, the surrounding page context would be able to just try again after a bit without manual intervention.
These are caused by the Toolserver. No idea why or what, but it seems related to their webserver setup. This appears at random on other tools of mine as well. As if some service is syncing files and their gone for a split second. Seems it was never reported yet at Toolserver Jira. Reported as https://jira.toolserver.org/browse/TS-1182. Although it's definitely a valid bug, it's not related to TestSwarm itself not to our instance on the Toolserver of TestSwarm at all. Marking INVALID.
Reopening, adding dep on bug 29550 (moving TestSwarm off of toolserver infrastructure).
Might be fixed with the new server so? :-)
Indeed, these were Toolserver specific and we're not using that anymore.
