techrecovery | risk of serious injury... (Reply)

days like today really ought to be banned. i have spent the entire day working with one customer, on one ticket. it is kind of frustrating to spend my entire day doing this when i have a list of stuff i need to get done both in regards to other tickets as well as the lab that i partly admin.

first, let me explain. they use a product of ours, and since they pay for a support contract, it is in our best interest to work our hardest to resolve issues relating to the software. a few days ago, they were doing maintenance on the system, cleaning up after a few employees left etc... so they removed the old users from the system. however, they didn't bother to check what those users did or were responsible for. one of the users was set to have their home directory in the same path as the database for our product (dumb dumb dumb dumb dumb). so when they removed the user, and ensured that the "remove users' home directory" was also checked, they removed the database, and subsequently, all files within it. and whala! our software no longer works. why? because it can't find the specified database!

well it took me about 1/2 day to come to the above scenario because i didn't think that it was a feasible situation, but turns out, it was. and who am i to comment on their configuration. i'm sure some admin had a good reason someplace along the line... so yesterday i made the recommendation that they restore the database files from a backup prior to user deletion and then to let me know how they make out. because, well honestly, it wasn't our fault, nor is it my job to be a UNIX admin for this guy who was new to HP-UX (though, honestly, it's no fault of his own, he was thrown into it). well they restored the DB but it wasn't the right copy. it turned out to be a newly initialized DB so they must have run the "dbinit" utility we have....

so today we asked them to restore all files associated with our software, an entirely new copy of the DB directory, and the /etc/passwd file to restore the two removed users. this way they'd be back where they were before all this happened. made them stop the server services, then restarted them after the restore was finished. test the connectivity via the client, and woah! it worked... [shaking head]

so here i am. i've now spent more than a day and 1/2 troubleshooting this one ticket which in the end, was a result of their own error. i can't tell you how frustrating this one scenario has been. UNIX admin who is unfamiliar with HP-UX, uninformed end users who don't know the name of the client software they're using, and a cloudy picture of what happened in the first place...

i'm tired now. i'm going to update the ticket, make a note to follow up with them on Monday and go home. i hope to be thoroughly sloshed in a matter of hours as a result of my totally craptastic day today.