nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Server reliability

Veli Voipio

Looks like Logos servers were down for 6 hours, without any notice, so I assume it was a crash.

As I understand FL has one server somewhere in NW USA. I suggest some backup servers, perhaps one in Finland and one in Australia, in order to have maximal geographical distance between.

Find more posts tagged with

Comments

Bruce Dunning

Here is what Faithlife Status is reporting. It apparently has to do with a switch upgrade - http://status.faithlife.com/incidents/ndplfjcfcgkg

Resolved - All Faithlife sites and services should be fully functional at this point. We encountered a series of hiccups within our network after upgrading one of the switches that have since been fixed. Stay tuned for a more detailed post-mortem.
Jan 28, 03:59 PST

Investigating - Our switch upgrade did not go smoothly and we are working to get things rolled back and back to normal. We'll keep you posted.

Lee

https://community.logos.com/discussion/comment/799584#Comment_799584

Veli

Also if you want you can get on the list to receive e-mail notice when Logos servers do have problems. I am on the list and i had 4 e-mails today about the issue.

Lee

Veli Voipio

https://community.logos.com/discussion/comment/799598#Comment_799598

That's good, but I am also thinking that it is good to have a service where one bad switch does not stop the whole system.

Mark Barnes

https://community.logos.com/discussion/comment/799602#Comment_799602

That's good, but I am also thinking that it is good to have a service where one bad switch does not stop the whole system.

It wasn't a bad switch, but a failed upgrade.

They'd planned downtime to replace the switch, but the new switch didn't work properly. They tried for a bit to get it to work, before giving up and replacing the old one.

delete12066188

https://community.logos.com/discussion/comment/799609#Comment_799609

Never touch a running System [;)] Things like this happend it's the Price we have to pay for having Internet and PC

Veli Voipio

https://community.logos.com/discussion/comment/799609#Comment_799609

It wasn't a bad switch, but a failed upgrade.

I mean if they had a system with parallel servers far away from each other, then it is much more reliable, whatever the reason for one server to go down of if some part of the Internet network fails.

In that case it still has two servers working. I mean from the user's point of view everything works, they are not interested why something is not working the just want to have something that works [H]

In many cases, like the the mobile phone networks, the upgrades and hardware replacements are done sometime 1 to 3 am. when there a few users and then they have some time to revert if something goes wrong. But Logos is already global, and there always some users that might be online.

JohnB

https://community.logos.com/discussion/comment/799598#Comment_799598

Veli

Also if you want you can get on the list to receive e-mail notice when Logos servers do have problems. I am on the list and i had 4 e-mails today about the issue.

Lee

I was informed by that same e-mail system precisely one week prior to the planned maintenance that it was to take place and when.

then one hour before

then when it was about to.

then one hour later that things had gone pear shaped.

I really don't see what more they could have done.

Matthew C Jones

https://community.logos.com/discussion/comment/799661#Comment_799661

I really don't see what more they could have done.

Amazon does not go down for 6 hours.

Parallel servers would have avoided this problem. A second new switch in hand would have lessened the length of the downtime.

This is the price of building castles in the cloud.

Simon’s Brother

https://community.logos.com/discussion/comment/799616#Comment_799616

the upgrades and hardware replacements are done sometime 1 to 3 am. when there a few users and then they have some time to revert if something goes wrong

One problem with this thinking ... with Logos users world wide we will all not be in the 1 to 3 am time zone at the same time there is never going to be a convenient time for all users.

If you sign up for the notification service at http://status.faithlife.com/ you will get pre-warning of scheduled downtime.

Networks are not my area so I'll leave comments on parallel setups etc to those of you who do know about those things and if there is better ways of handling these scenarios I hope Logos is looking into these solutions but I assume these solutions will come at a cost that ultimately we need to pay for as uses of the service somewhere along the line.

MJ. Smith

Looks like Logos servers were down for 6 hours, without any notice, so I assume it was a crash.

No it was scheduled maintenance for an upgrade. http://status.faithlife.com/ will always give you such information. It did take longer than the original notice indicated they expected.

MJ. Smith

https://community.logos.com/discussion/comment/799616#Comment_799616

I mean if they had a system with parallel servers far away from each other,

They do have a backup site about 100 miles south of their main site. I don't know why it was necessary to interrupt service. For their time zone, they aimed at 10:30-12:00 but things went awry ...but given that my television goes out for a few minutes every night a little after 2:00 am I have no trouble uctting the much smaller Faithlife a bit of slack.