A Google Cloud support engineer solves a tough DNS case

Post by enedil » Wed, May 20, 2020, 1:00 AM UTC

DethNinja wrote: post #141

mcguire wrote: post #129

" When sk_rcvbuf gets close to 2^31, adding the size of the packet can cause an integer overflow. And since it’s an int it becomes a negative number, therefore the condition is true when it should be false (for more, also check out this discussion of signed magnitude representation). " And this is why you don't generally use signed numbers in systems code, unless you specifically need negative numbers. And why you gr…

Signed or unsigned doesn’t matter much both can overflow. I actually started putting assertion checks about overflow issues almost everywhere but it requires great discipline. I wonder if there is a better solution available.

Post by jagannathtech » Wed, May 20, 2020, 1:40 AM UTC

jiggawatts wrote: post #192

grawlinson wrote: post #189

Earlier quoted context omitted.

One of my computers has been aptly named 'THESEUS' due to what was replaced on it. By the time it was repaired to an acceptable level, the only original component remaining was the chassis.

Heh... that reminds me of my favourite troubleshooting story! We had a customer with regularly failing tape backups. CRC errors, verify pass failures, even failed writes, and so forth. We replaced the tapes with new ones. Same issues. We replaced the tape drive with a new one. Still the same problems. We replaced the internal ribbon cable and the SCSI controller. No luck. Firmware flashed everything. Didn't help. New…

Post by JackeJR » Wed, May 20, 2020, 1:57 AM UTC

grawlinson wrote: post #189

jeffbee wrote: post #95

Earlier quoted context omitted.

It is indeed folly to assume that cattle have no identity. There's an internally famous video inside Google in which a particular well-known machine was unracked, dragged out into a field, and ceremonially smashed to pieces by some hardware techs. Sometimes a machine just takes an arrow to the knee and it's never the same again. Then there are all the uncontrolled or unrecorded differences between machines: the ones…

One of my computers has been aptly named 'THESEUS' due to what was replaced on it. By the time it was repaired to an acceptable level, the only original component remaining was the chassis.

Post by perfect_wave » Wed, May 20, 2020, 2:21 AM UTC

ehsankia wrote: post #146

andyljones wrote: post #66

Earlier quoted context omitted.

Here you go: https://cloud.google.com/support#support-plans $250/month/dev is the minimal for phone calls on technical issues, $150k + 4% of GCP spend for 'come running' support. There're more details here https://cloud.google.com/support/docs/procedures#additional_... though they use the old names for the support tiers.

It seems like the blog post talks about a written case report, which the 100$ tier has access to, albeit with 4 hour first response instead of 1 hour. So it is possible that you could get your case escalated to such an in depth debugging with that tier?

Post by mav3rick » Wed, May 20, 2020, 2:26 AM UTC

floatingatoll wrote: post #202

mav3rick wrote: post #196

Earlier quoted context omitted.

Yes I qualified it at the end why the former is the correct solution and not the latter.

I figured that "in reality this would have worked too" is a easy phrase to misinterpret as "this would have worked" (assuming they miss the context a paragraph later), and so the reply helps ensure that others do not misread it as I initially did.