On Monday, November 15th, 2021 at 21:53, Scott Bennett bennett@sdf.org wrote:
The authority relays have exhibited very unstable behavior
for, I think, nearly two years now, as I have pointed out several
times on this list. For about three months now, there have been
signs of improvement, but the situation is still not terribly
stable or predictable. The two flags that appear to me to have
been most affected are HSDir and Stable, but Fast has also come
and gone occasionally since the instability began. I don't run
a Guard relay, so I don't know about that flag's stability, but
I would imagine that it would be affected by instability in the
Stable and Fast flags. All three flags can come and go, seemingly
at random, without regard to uptime of a relay or stability of its
own characteristics and network connections.
Yes I have also lost the Guard and HSDir flags. I agree with you that these seem to be predicated on gaining the Stable flag, which is my focus for this thread. Basically, most of the DAs do not believe my relay's reported uptime, which at the time of writing is over 12 days. I would like to understand why this is happening. I suspect network issues (routing, peering or socket resources) are at play here but instead of spending time speculating I would prefer if the DA operators could assist me with this. The DAs which are witholding the flag are tor26, gabelmoo, dannenberg, maatuska and bastet.
I also agree with you that there appear to be wider problems with the consensus health overall, but to me the issues have increased noticably in the past few months.
‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On Tuesday, November 16th, 2021 at 01:53, dg26 dg26@cock.li wrote:
How much RAM are you running Tor in your instance? Probably not enough
RAM and getting the TCP Out of memory.
2GB of RAM overall and over 500MB still available. Not a RAM issue.
Would you mind sharing the results of `netstat -st` ?
I'd rather not share those kinds of outputs publicly.