Database Crash affecting our rankings

[tdawson09]

Here is one for you. We had a database crash, and was down for an hour and 15 minutes, on Dec 18th. During this time our system automatically sent out a 301 redirect to our product home page. (iboats.com/boat-parts-accessories/dm/)

On Dec 24th I noticed a huge ranking drop on a lot of our keywords, and some keywords that were serving up the redirect page from our database crash, in place of its normally ranked page.

So my assumption is that Google was crawling our site during the database crash and caching the redirect pages in place of our regularly ranked pages. So when a search query is made that would normally display our ranked page, it is now displaying the redirect page from the crawled cache.

And on other keywords where we normally would rank for, Google no longer has that page cached, so it drops it from ranking.

My question: Does this assumption sound accurate? If so, I'm assuming over time that our past ranked pages will show up once again after Google recrawls those pages and saves them in their cache.

I have several sites that were affected with this:

boatcovers.iboats.com

boatpropellers.iboats.com

iboats.com

Thanks in advance for your input

[dawnieando]

Glad to hear you appear to be back on the right track

[tdawson09]

Update:

After unblocking Google, almost all of our rankings are back up, and in some cases a little improved. I was surprised how quickly Google recrawled and ranked the pages that dropped.

I did do a Fetch and submit on all my websites and pages. I'm not sure how Fetches are attended to by Google, but it appears they are.

I have an IT guy who doesn't feel a Fetch makes any difference. He says Google will crawl when they want to and probably does not pay attention to the Fetch request.

[tdawson09]

I wanted to get back to report the cause of our problem with a lot of our rankings dropping. We originally thought it was due to a database crash. And that still could be affecting some things.

We moved our servers to a new location. A couple of weeks ago, we switched over to some new IP addresses. When we did so, a program started blocking Google's IP addresses because they were hitting our servers with a lot of hits, and put those IP addresses on a blocked list. So our pages on some of the websites stopped getting crawled, and thus rankings dropped. We discovered the problem when exploring our Webmaster Tools Crawl Stats and noticed that the "kilobytes downloaded per day" had flat-lined. Doing a fetch on selected pages we could see that it was not fetching the page. So after unblocking those IP addresses, the pages got crawled, our crawl stats went back up, and our rankings started going back to where they were. We still have some rankings that have not bounced back, but assume they will as new crawls re-index those pages.

A lesson learned for me was to pay attention to the crawl stats. Any drastic variations could be a sign of a problem.

Brad
iboats.com

[dawnieando]

I'm pretty sure that GWT is a little vague in  terms of times that Googlebot visited with just the crawlrate shown on a day by day basis and there's some lag between now and when it happened.  CPanel will show when spiders visited in latest visitors from memory, however I think that rather than looking back it would be better to simply ensure that this doesn't happen again with the 302 rather than the 301.  Maybe refreshing of your sitemap submissions in GWT would help to get some crawling going on with a ping or two to get Googlebot back soon.

[williammarlow]

Also, can't we use GWT to find out the exact times or at least approximate times of when Google crawled a site? Or can we not find that information so precisely?  In this case, it does seem that the 301 vs. 302 is the key problem, but also the fact that the system for getting rid of the 301 didn't work out properly.

[dawnieando]

It seems extreme that Google would react within an hour and 15 mins but who knows nowadays.  Keep an eye on your GWT account and hopefully if you have a decent crawl rate things should remedy themselves.  Stephen Salstrand made a good point.  302's rather than 301's should help.  It's great to have something in place like a redirection for user experience in case of database failure, however, as Stephen said, it should be made clear to Google that it's only temporary.

[tdawson09]

Investigating it further, we have some permanent redirects that are still hanging. I've got my IT guys working on it as we speaking.

I've got to run, but I would like to revisit this soon. Once we have determined the problem, I will post it here.

Thanks for your opinions. They have helped.

[XNUMERIK]

Your assumptions are globally right. However, I have never seen Google react that quickly to a 301 redirect or even a crash, I always tell my customers "if you have a temporary issue with 404 and 500 erros, don't worry about it too much, Google won't react unless it is a permanent issue that lasts over three or four crawls".

This is why I would also investigate a penalty that just happened to take effect shortly after the crash. To rule that out, could you please tell us if some of your major keywords are still ranking well?

I don't think you sould worry too much about that, just wait a few days and see if some pages recover.

[tdawson09]

Good point on the 301's. I'll look into that.

Thanks, for your input!

[SteveTheWoodsman]

I think your assumptions are on par.  However, make sure to review that the 301 has been removed.  Also, if you still want that auto redirect to happen on a crash, make it a 302 and not a 301 as a 301 will suck a lot of life out of your site (since SEs see it as permanent).