Hello, So recently I rebooted my relay using the killall tor -hup command (while I was trying to mae some changes to my torrc file). My relays uptime was not changed at all by this, having a current uptime of about 6 days. However, I noticed today at http://torstatus.blutmagie.de/router_detail.php?FP=db1af6477bb276b6ea5e72132... My relay lost the stable flag (which it had had for the last about two months). I know not to worry about exactly what flags I have all the time, but it raises the question of what eactly the killall -hup command does? Thanks.
Keifer Bly:
it raises the question of what eactly the killall -hup command does?
it sends the HUP signal https://www.torproject.org/docs/tor-manual.html.en#SIGHUP
SIGHUP The signal instructs Tor to reload its configuration (including closing and reopening logs), and kill and restart its helper processes if applicable.
-- https://twitter.com/nusenu_ https://mastodon.social/@nusenu
Yes, but I randomly lost the stable flag last night after using the killall -hup command twice in one day. I did it strange that randomly happened even though my uptime did not change after doing this. Cheers. From: nusenu Sent: Tuesday, July 3, 2018 12:46 PM To: tor-relays@lists.torproject.org Subject: Re: [tor-relays] killall -hup command Keifer Bly:
it raises the question of what eactly the killall -hup command does?
it sends the HUP signal https://www.torproject.org/docs/tor-manual.html.en#SIGHUP
SIGHUP The signal instructs Tor to reload its configuration (including closing and reopening logs), and kill and restart its helper processes if applicable.
-- https://twitter.com/nusenu_ https://mastodon.social/@nusenu
On Tue, Jul 03, 2018 at 04:23:27PM -0700, Keifer Bly wrote:
Yes, but I randomly lost the stable flag last night after using the killall -hup command twice in one day. I did it strange that randomly happened even though my uptime did not change after doing this.
The Stable flag, like Guard and Fast, is relative: it has to do with how your relay compares to the other relays in the network at the time. So you can leave your relay alone and yet it can still add or lose various flags, depending on what happens with *other* relays. For example, maybe some historically more stable relays came back online. This will be especially true if your relay is already on the borderline of whether directory authorities think it should get one of the flags, since a little nudge in one direction or another can make a difference. My guess is that the stable flag assignment had nothing to do with sending the hup signal, and it was just a coincidence. --Roger
The 'kill' commands send various signals to processes, as defined in signals.h (see https://linux.die.net/man/7/signal). -hup sends SIGHUP to the process(es), short for hangup. This historically was for when a serial connection was dropped, and the process needed to close / take action accordingly. For daemon processes, and more commonly with modern software, the response to SIGHUP is to re-read configuration files and restart the process. It is likely the case that you are losing the stable flag because Tor drops existing connections in the process of restarting and reloading configuration when it receives SIGHUP. Someone more familiar with Tor internals might be able to confirm what exactly is happening there. -MrDetonia -------- Original Message -------- On 3 Jul 2018, 04:28, Keifer Bly wrote:
Hello,
So recently I rebooted my relay using the killall tor -hup command (while I was trying to mae some changes to my torrc file). My relays uptime was not changed at all by this, having a current uptime of about 6 days. However, I noticed today at http://torstatus.blutmagie.de/router_detail.php?FP=db1af6477bb276b6ea5e72132...
My relay lost the stable flag (which it had had for the last about two months). I know not to worry about exactly what flags I have all the time, but it raises the question of what eactly the killall -hup command does?
Thanks.
I already got the stable flag back this afternoon thanks. Just thought it was a thing to note. From: Zac Sent: Tuesday, July 3, 2018 11:57 PM To: tor-relays@lists.torproject.org Subject: Re: [tor-relays] killall -hup command The 'kill' commands send various signals to processes, as defined in signals.h (see https://linux.die.net/man/7/signal). -hup sends SIGHUP to the process(es), short for hangup. This historically was for when a serial connection was dropped, and the process needed to close / take action accordingly. For daemon processes, and more commonly with modern software, the response to SIGHUP is to re-read configuration files and restart the process. It is likely the case that you are losing the stable flag because Tor drops existing connections in the process of restarting and reloading configuration when it receives SIGHUP. Someone more familiar with Tor internals might be able to confirm what exactly is happening there. -MrDetonia -------- Original Message -------- On 3 Jul 2018, 04:28, Keifer Bly < keifer.bly@gmail.com> wrote: Hello, So recently I rebooted my relay using the killall tor -hup command (while I was trying to mae some changes to my torrc file). My relays uptime was not changed at all by this, having a current uptime of about 6 days. However, I noticed today at http://torstatus.blutmagie.de/router_detail.php?FP=db1af6477bb276b6ea5e72132... My relay lost the stable flag (which it had had for the last about two months). I know not to worry about exactly what flags I have all the time, but it raises the question of what eactly the killall -hup command does? Thanks.
participants (4)
-
Keifer Bly -
nusenu -
Roger Dingledine -
Zac