[illumos-gate] 14986 SMB Shares appear to be rename-able from Clients -- Alexander Stetsenko <astetsenko⊙rc>

mmm, I do love how gnu make -j is fork bomb and how latest and greatest linux is dead from that....

is delphix.run still needed in zfstests?

didn't they switch to OpenZFS [and linux]?

We can probably nuke it unless we wanna keep it for historical reasons (which may be reason enough to NOT nuke it...).

it'll be waiting there in the history if we do want to resurrect it

In that case, feel free to nuke it?

for history, we do have git anyhow.

[illumos-gate] 14969 Improve elfcap.chk CPPFLAGS -- Robert Mustacchi <rm⊙fo>

danmcd: (since you'd probably know offhand if anyone).. is there an easy way to tell how long a TCP socket has been sitting in a given state?

(e.g. i'm seeing about 6000 sockets all in SYN_SENT -- just wondering how long they've been sitting there)

I don't know off the top of my head.

There probably is a timeout somewhere in conn_t or tcp_t, but I'll be damned if I know what it is immediately w/o source-diving.

ok

(also: netstat -u doesn't show sockets in SYN_SENT apparently)

that feels like it might be a bug

I dunno if we track anything like hrtime of last segment for a given conn, but if we did that would be nice to put in connstat

if this is reproducible you could use dtrace and the TCP state change probes

TCP/IP doesn't use as much hrtime as it should. A lot of it is still lbolt. (I know IPsec needs some SERIOUS rewhacking there e.g.)

yeah, it would be nice to base that all on hrtime

(eventually)

i suspect what's happening is that this thing is immediately spawning several hundred connections to the same S3 IP from behind a nat and the firewall involved is displeased

and is just going 'umm no' and dropping the packets on the floor

should be easy enough to confirm with snoop or `tcp:::send`

r1mikey ping? Are you still working on your Arm fork?

jclulow gerrit does not like me: error: remote unpack failed: error Missing tree 9acf9ad2a30823c04a55d2cd30592403745dbe79

fatal: Unpack error, check server log

jclulow ! [remote rejected] HEAD -> refs/for/master (n/a (unpacker error))

you're pushing something that doesn't reach a changeset we have, I think?

previous one is 663b59f48c396b41650fffe9e2f20a85e9710c85

let me rebase, I had one missing...

hm, now it did work. so it wants me to be on latest.

[illumos-gate] 14961 validate_pkg crashes with bad manifest -- Andy Fiddaman <illumos⊙fn>