[ SmokeDetector | MS ] Bad keyword in body, blacklisted website in body, link at end of body, link following arrow in body, pattern-matching website in body, +1 more: Can you relate to this situation? by Bodem 19 on superuser.com
[ SmokeDetector | MS ] Bad keyword in body, blacklisted website in body, link at end of body, link following arrow in body, pattern-matching website in body, +1 more: Can you relate to this situation? by Bodem 19 on superuser.com
@Glorfindel Lada was the name for export, in Russia they are known as Жигули. I'm not surprised you didn't see them; if one can afford to live in Moscow or StP, one can probably afford a better car. Also, imported cars (mostly 2nd hand) took over the Russian automarket decades ago, somewhat countered by punitive tariffs in more recent years.
> AWS Free Tier includes 750 hours of Linux and Windows t2.micro instances each month for one year. To stay within the Free Tier, use only EC2 Micro instances.
This seems like a pretty crippling issue that we've seen a few times. How large is the disk currently? The storage costs in EBS are pretty tiny. Surely we can smash a bigger disk on it and then just let it take longer for this to happen. At least that would decrease the frequency of this.
Also we’ve offered him donations before, but he can’t take them because he uses that server for things other than MS and it would make his accountant confused if donations were included
@ByteCommander I've got no idea what the problem is. What I'm trying to convey, evidently sub-optimally, is that whenever MS has a perceived disk space issue, the instances all appear as if they're not working correctly when, in reality, they cannot communicate with MS to indicate that they are working correctly.
@Undo This issue takes some people hours to debug if they don't know of the concept of inodes (df -i and du --inodes /* are useful in these situations)
I find it quicker and easier to drill down using the following command:
$ sudo du -s --inodes * | sort -rn
170202 var
157325 opt
103134 usr
53383 tmp
<snip>
You can then go in to var for example and see what the big inode using directories are in there.
@Henders no, I mean, I understood what you wanted to say, but the wording you used made me think of something else... someone not to be explained here...
@Undo You migth be able to use one of the solutions on this page: stackoverflow.com/questions/347620/… I think one of the perl/python scripts migth be your best on your box, because most people use du --inodes
Or you can follow it from the principle that every file equals 1 inode, and search for directories with thousands of files
Most members of Mi Yodeya are going to be offline all of Friday and Thursday, and there's this troll around, so please remember to pay some extra attention there for the next two days. Thanks
@GrantGarrison it's a general detection pattern, not something we can selectively whitelist without some heavy duty restructure in the detection pattern
@GrantGarrison No you don't. Multiple urls in title still count as one reason. Url in body only counts if it looks spammy (matches a pattern, or is blacklisted)
And if the reason weight is that high, then it means it's a very good reason
@GrantGarrison True. But I'm not addressing that, I'm addressing the fact you said "Throw in a few more links and we have a fp autoflag", which is not true.
@GrantGarrison you shouldn't judge the individual reasons too heavily on their fps. With reporting, it's better to have a few too many reports, as long as they are not flagged. Note that this is entirely different to autoflagging.
I find it quicker and easier to drill down using the following command:
$ sudo du -s --inodes * | sort -rn
170202 var
157325 opt
103134 usr
53383 tmp
<snip>
You can then go in to var for example and see what the big inode using directories are in there.
