Has anyone else noticed PQs returning out of date information. This morning I planned to hunt a few caches in Atlanta so I set up a new PQ centered on downtown and submitted it. Since it was a new PQ, it ran quickly and was in my in box in a few minutes. One of the caches I was interested in had several log entries yesterday and the coords for the cache were changed yesterday. However, the PQ did not include any of yesterdays log entries or the change in coords.

I understand that PQs are run on a dedicated machine and I suppose that accounts for the lack of real time PQs. Has anyone else noticed this lag?

I have, now that you mention it. I thought something was kinda goofy. It seems that the data is about 24-36 hours out of sync.. like maybe the PQ machine isn't pulling from the "master" db.

I haven't bothered to go to the groundspeak forums yet to read about it. They'll either figure it out and fix it or it's the "new standard." Probably the latter..

I have noticed the PQs I have scheduled for Friday don't arrive until Saturday. Wonder if we are going go through another round of PQ purging?

Here's the groundspeak thread.

Looks like it's been fixed.

Mertat (or anybody else), do you remember the gist of the "fix"? I didn't read the resolution before gc.com went poof.

I believe Jeremy said the computer that runs the read-only copy of the database which is used by the PQ generator had problems so they brought up another computer and placed a copy of the database on it.

That sounds right. I guess "fix" wasn't really the best word to use. They had a workaround until they could actually get it fixed.

I just got our "My Finds" PQ and all the info was there and correct...so the workaround is working.

Now for them to get the forums up......

quote:

Originally posted by AllenLacy:
I believe Jeremy said the computer that runs the read-only copy of the database which is used by the PQ generator had problems so they brought up another computer and placed a copy of the database on it.

The *computer* had problems that prevented it from having the most recent data?

Uh, ok. Trout, Denmark, and olefactory impoliteness. Regardless, as long as it's working correclty now.

Well the message said

quote:

We had a database failure in one of the secondary (Read-only) databases.

So it wasn't the computer, but the actual secondary database. I could say something about using M$ database software in a high demand environment but I won't.

quote:

I could say something about using M$ database software in a high demand environment but I won't.

I've worked in environments with Oracle, Sybase, iSQL, and MSSQL in high-demand production environments and all have gone down at least once that required multiple-hour recovery and usually with vendor help.

Usually MSSQL installs failing are the same as with the OS, they are just thrown together by people who don't know what they are doing (usually because MSSQL is the lowest cost).

My current job is at a place that has over 150 servers housing a few thousands MSSQL databases and hopefully they will all continue to work well (or I might lose my job).

reepicheep

Well the good news was the primary database was ok; the bad news was the secondary database was messed up and they didn't know it until the users complained, repeatedly.

One thing that appears to be a weakness of MS SQL Server is that as far as I can tell it can’t do hot backups, i.e. backup the database while it is in use(1). Jeremy mentioned they shut the primary database down once a week, to back it up to tape and once a month to re-index. 15 years ago in a shop using Oracle we had to stay up 2 nines (99% of the time), so we did hot backups to tape. My current position our goal is to keep our DB2 database up three nines (99.9%) which allows us to be down 43 minutes a month. So we do hot backups to tape and keep a redundant copy on a different box. I have talked with the DBA of a telecommunications company using Oracle and their goal was 5 nines (99.999%) which allows them to be down 5 minutes and 15 seconds a year. So they do hot backups and have 2 redundant copies with automatic switch over. Never seen a MS SQL Server that would stay up that much.

(1) I have seen 3rd party software which allows hot backups of MS SQL Server.

As an IT old-timer I'm pretty skeptical about any company or product. And products seem to get buggier every year. But as much as I despise Larry Ellison's arrogance, I have to admit that Oracle is a dang good product. I oversee half a dozen Oracle databases and I've seen them survive some pretty catastrophic scenarios. Their hot backups are just short of magic. And in their "Backup and Recovery" class the teacher keeps crashing your database (e.g., deleting files) and you have to keep recovering it before you get to go home.

Good points by all.
I think we scared all non-IT folks off

reepicheep

Also I have to admit that Jeremy + all the groundspeak personnel provide very good service at a very cheap price $30/year, $3/month, or most features for free, plus a little ad money.

Now if I was paying $100+ a year, I would expect them to be up 99.99% of the time. But for the amount of money they take in, the service level provided is fine.

Me thinks it is best to remember that geocaching is a game. The data here is of value but if my pocket query isn't up to date to today that is an inconvenience. For the game, it is a delay perhaps but lives are not in jeopardy, personal ID data isn't compromized and nobody loses his/her financial shirt because vital transactions got lost. We all want the data there and correct but it is just a game. It got fixed and full functionality got restored, so maybe folks didn't get a FTF because a PQ didn't have the latest info for a few days. Although inconvenient, the sky isn't falling.

Personally, I think GC.COM does a great job at keeping our game functional. They work at giving us accurate and timely data so we can enjoy our game. I give them a thumbs up.

Just my $0.02 worth.