47

> Customer logs Jira ticket claiming app is not working
< I restart the app, investigate and explain tht their server has issues
ø Client closes the ticket as Resolved
-- a couple of days pass by ---

> Customer logs Jira ticket claiming app is not working
< I restart the app, investigate and explain tht their server has issues
ø Client closes the ticket as Resolved
-- a couple of days pass by ---

> Customer logs Jira ticket claiming app is not working
< I restart the app, investigate and explain tht their server has issues
ø Client closes the ticket as Resolved
-- a couple of days pass by ---
<...>

< I log a JIRA ticket explaining what and how is wrong with the server with suggestions how to fix the problem so the app will not crash any longer (client own the server, has his own sysadmins -- I don't even had permissions to open syslog.. had to hack dmesg on their PROD server to pin-point the issue)
> no reaction from customer for weeks. I ping the ticket

× app crashes again

> no reaction from customer for weeks. I ping the ticket
> customer leaves a comment that their sysadmins are looking at it trying to figure out what might be wrong (ignoring what I wrote in ticket's description??? srsly?)

× app crashes again
< I post detail investigation details: snips from logs, screenshots, everything with crystal clear explanations.

> no reaction for weeks
......

well that's fun..

Comments
  • 8
    Leave them alone. Why do you even care that much? Let them fly on their own, if they crash then it's not a problem
  • 1
    An app that crashes, is an app that crashes. Implementing a proper catch on not able to connect to the depending server and ensuring that the app does not crash, even if the server is down is a way to totally isolate the issue. Both client and server have shortcomings. Placing the issue totally on serverside is not really fair😊
  • 3
    @p0s1x dmesg ir filled with segfaultd and hdd timeouts that are freezing the whole app server along with app jvm, effectively making the app unresponsive for up to half an hour. Watchdogs catch that freeze [lack of heartbeat ack], nuke the jvm and restart it.

    Now how do you handle faulty dimms and burned down vmware hypervisor's hdds in the app?
  • 0
    @netikras Ouch that indeed seems a big issue, none the less, any client should poperly handle the absence of the server ( that is how i would isolate the problem, therefore having no issues on the client (in an ideal world :)) . I hope the issue with the server hopefully gets tackled, good luck mate!
  • 0
    Ahh now it’s clear to me, the app runs on the mentioned server, yeah thats mess :) Sorry i provided an reaction bases on the information provided.
  • 1
    @p0s1x yeah, if it was a client-server app model it could be handled easily. With a nice message 'we are sorry our servers have bugs. A dozen of highly trained tits are hunting them now as you read this message' :) but when it's a webapp w/ a web interface... It's a little bit harder :)
Add Comment