Thief

joined 2 years ago
 

I have decided the best way to resolve the issues of disk space on the server is to change VPS providers. The server setup has to continue to be extremely cheap and incur no cost to the community, and the current provider does nto have plans that include the required disk space at a fair price point.

As a result I will need to change providers. The current VPS is : 8GB, 4CPU, 180GB disk.

I wish to change providers to this setup: 16GB, 6VPU, 400GB disk.

The cost is similar but as you can see I get more resources from changing the provider. This does mean I have to take everything offline, copy everything to a new VPS over the internet and then reconfigure everything on a new VPS. As a result I am selecting 14,15 and 16 February to do this. The server will either be down or only semi accessible on these days.

-P

 

Unfortunately after over a 18 months or so of service, the server requires some maintenance. The disk space on the server keeps running out and causing issues. The server was setup very cheaply so this is expected, but it does mean I have to take the server offline at some point and fiddle around with the database, clean it up and a host of other annoying things that will take several hours to complete.

Im not sure of a good time to do this but since the server has been dropping offline every now and again and requiring I log in and bring it back up for the last week, its probably going to happen soon.

Once the maintenance is done we should be good but just want to let you know to expect some disruption.

-P

[–] [email protected] 16 points 2 months ago

Imagine they tell her they are swifties… she would have used to have mad love for them but now it would be bad blood.

 

Upgraded to 0.19.7 today to keep up to date with Lemmy.

[–] [email protected] 2 points 7 months ago

Upgrading is always a nightmare but once its done generally lemmy improves bit by bit :)

 

Upgrade did actually complete. pictrs had some issues but I think I fixed them (hopefully).

Any issues let me know.

 

Lemmy once again needs upgrading. This time all components - pictrs, postgres and lemmy itself. This will mean that at least 1 of these (fragile) things will almost certainly break and leave the server down for some time. I will be investigating how best to unbreak the server over the next couple of weeks and post when we intend to do the update that inevitably will destroy the server in some way. If you use the server daily you may wish to create a secondary account so that you are not too disrupted when the update happens. Unfortunately upgrading lemmy is not as easy as one might hope so I cant tell you exactly what will happen ahead of time. I will post again when we have a general idea what to do with an intended date.

[–] [email protected] 3 points 8 months ago

Just to let you have an update, we are caught up with all instances except lemmy.world which is still 3 days behind. You can monitor or track the federation progress here: https://grafana.lem.rocks/d/bdid38k9p0t1cf/federation-health-single-instance-overview?orgId=1&var-instance=lemmy.world&var-remote_instance=lemmy.myserv.one&from=now-12h&to=now

[–] [email protected] 8 points 8 months ago

Either a local SMTP server (less used) or an external service (more common). The SMTP is configurable but I believe most used option is ssl smtp over port 587.

[–] [email protected] 2 points 8 months ago

All good, sorry for the downtime :)

[–] [email protected] 3 points 8 months ago (3 children)

Federation is returning to normal and should be working again fully by tomorrow. There is some catch up to do.

[–] [email protected] 3 points 8 months ago

Thank you, I am busy looking at the federation issue right now and hope to resolve it soon. This instance is just a single server so if its turned off then the impact is total, but in terms of the datacentre the entire site was down and power was slowly restored to parts of it and is still ongoing, we are luckily back online but some parts are still down and damaged. The site is 1515 round table drive Dallas TX.

I also purged the post you reported.

 

Hi there,

Unfortunately the site was down for about 2 days over the 26-27 May this month.

This was due to the datacentre we use being affected by a storm/tornado: (https://apnews.com/article/tornadoes-weather-texas-oklahoma-93001e1d81f120d0d55c8dfa24938250) which took out power to the facility for just under 2 days. At no time was there any risk of loss to the server due to off site backups, however we did have to wait for the centre to come back online.

While this server is run extremely cheaply, at just $20 USD per month (approx) which includes paying for the compute, domain, email, and off site backup as well as object storage, it does mean that there is not redundancy when a site wide outage occurs. The low cost means that there is no risk of this server ever being unable to meet its costs and be paid for, but it does mean we sometimes can have an outage if a major event happens such as what happened and a delay before being brought back online.

The decision for this was because this server is designed to help reduce load on the rest of the fediverse and does not store much local content so a couple days a year the server is out was deemed acceptable during the design phase.

If you have any questions or comments feel free to post and I will try my best to answer them :)

  • P
[–] [email protected] 7 points 10 months ago

nominative determinism

 

Upgrade to 0.19.3 was completed with no issues today (from 0.19.1)

[–] [email protected] 1 points 1 year ago (2 children)

I mean so is control d a set and forget. Even more so because they manage any changes and workarounds the companies come up with. Its basically designed for this. Having tried both options its my preferred option.

[–] [email protected] 1 points 1 year ago

Set your country to Albania, Moldova, or Myanmar and you wont get ads on youtube.

 

Just want to give everyone an apology for the downtime last few days. Its unfortunate that we were down around 18 - 20 hours in total and rest assured I have made improvements based in what happened to reduce this going forward. Its been 24h now since the upgrade to 0.19.1 and everything has remained stable so far.

Hopefully thats the end of it. Merry Christmas and a Happy New Year to you all :)

-P

 

Today on 27 October 2023 lemmy.myserv.one reached 98.5% uptime. We are also in the top 700 most reliable lemmy servers now as measured by this site: https://lemmy.fediverse.observer/list

Next goal is top 500 servers and/or 99.5% uptime.

Thanks everyone:)

3
Racism (www.ctvnews.ca)
 

Reminder in light of recent events for members here to always conduct themselves appropriately on other instances and not make racist comments towards any race. Keep opinions clean, factual and constructive and be dignified in your responses.

I would prefer not to have to act on other admins reports that members from this server have been acting inappropriately.

Many thanks in advance for your considered and exemplary behaviour.

-T

 

Updated the server to 0.18.5 today so it is kept up to date and compliant with the rest of the lemmy fediverse.

 
view more: next ›