Rolling Restart for 1.25.5, Feb 3-5 2009

This group is very small with only a few members. Before the new server, Second Life Server 1.25.4.108489, I had no problems with the group loading. After the new server was installed, I had to ask the sim owner to reboot the sim. The group info loaded for a couple a days and now the group info is loading. If it is not the server and there is only a few members in the group, (19 in all), what else can it be? It does not matter either if it is during peak hours or not, group info will no load.

Second Life Server 1.25.4.108489

With 1.25, we are seeing one of the apache2 process occasionally go nuts. It's rare, but when it happens it can cause things to get bad on the host where it's running. If you're having problems such as inexplicable bad lag in a sim, contact support. They *may* be able to help.

I am not sure if the group loading problem is related to this or not. I have heard a few reports of it. Has anybody filed a JIRA on it? If so, what's the JIRA number?

The ticket number I filed is Ticket 4051-5764755 and it is listed as closed. I was informed to contact the sim owner for a restart of the region. The region was restarted and the groups loaded for a couple of days. Now the group info loading issue is still a problem.

The ticket number I filed is Ticket 4051-5764755 and it is listed as closed. I was informed to contact the sim owner for a restart of the region. The region was restarted and the groups loaded for a couple of days. Now the group info loading issue is still a problem.

I think this is what your wanting Prospero

https://jira.secondlife.com/browse/VWR-9655


I don't think this is a viewer side bug was only one i could find that had the group loading info in it

https://jira.secondlife.com/browse/SVC-1815

there is also that one which says about the affect the group loading causes.

The pilot rolling restart will begin shortly. (I was bouncing some of the central hosts that needed an upgrade in the last 1/2 hour.)

I'm at my island in-world, and will be for the next ~hour, if you want to grab a commemorative bear. I'll be out again during the full rolling restart over the next couple of days.

Because of a small goof I made, a handful of regions will get restarted twice (about 20). Sorry about that :|

How far along is the pilot restart?

Please show the update has started from the status page entry. Having to hunt for the status is annoying.

Thank you

Wow ... it's been hours since the 1.25.5 restart of the pilot regions and no drama? I'm impressed Prospero! Congrat's!

Chat and private messages come up repeatedly two or three times per instance.. with big lag! Status page says nothing about that at all. Havent this been noticed?

Thanks!

Prospero why do you guys keep rolling out this when it looks like it suits you. I think at least you could send out a gridwide warning like you used to do when i joined SL in 2006/2007. At that time you also had every 2nd wednesday where you closed the grid for residents, and then you did major gridupdates in 4 hours.

I talked to many people that actually want the wednesdays back instead of the updates that you do now, if i knew how to make a JIRA (or what it's called) so people could vote for it i would.

I'd have to agree with this. Every time there is a rolling restart the grid seems to take time to recover and the user experience is badly degraded. I would much rather have a 4 hour outage than 3 days of an unreliable SL. In my opinion it's better to have a usable SL instead of minimising downtime.

Er.. The old Shutdown-Wednesdays rarely finished on time and the grid was at least as likely as now to have issues afterwards. And if anything went horribly wrong, they had to shut the whole entire grid down again to revert it.

They would be insane (and likely fired) if they tried to go back to the old way. There's far, far less downtime now..

Why am I having problems connection to the simulator now?

The same thing used to happen: you'd have a 4 hour outage on Wednesday, that occasionally turned into a 12 hour outage, followed by instability on Thursday, and occasionally a second downtime on Thursday if they couldn't manage a rolling restart, and more instability on Friday.

I'd rather have 3 days of even serious instability (and some of these staged rollouts have been so mild I didn't even realize it had started until it was almost over) to "three day upgrade Wednesdays".

People who want to go back to 2005 or 2006 just don't remember how bad the good old days were.

The "3 days of instability due to a rolling restart" is a myth, by the way. Yes, there were login problems for a while for some today, but that had nothing to do with the pilot roll this morning.

Excuse, there can be I not in a theme, but nevertheless. On February, 3nd my friends had a bug on sim. All objects containing scripts, have been returned owners, and by this is rebooted. In a local chat it has been written that objects are placed in a folder "lost and found", but actually, any objects has not returned. They are very upset, as have lost set of the objects which do not have attributes of copying. Whether there can be it a consequence of reboot and how it to return the lost objects?

Do we know if it is odd or even regions that are going to be restarted today?

The first-half rolling restart has just begun. Odd-numbered hosts are being restarted today.

At 08:08, we're about 34% complete.

At 09:10am, we're about 79% complete.

At 9:49, we're done.