Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
elwoos
Master Smack Fu Yak Hacker
2052 Posts |
 Posted - 2004-11-01 : 14:10:09
|
| I've just been on annual leave for a couple of days at the end of last week. Strangely enough, things happen while I am away (co-incidence honest!!). Last year it was a major virus outbreak, this time it was the power to our production server! My friends in Network Admin apparently came to 'look at the problem' and declared that the cooling system that they installed a few weeks ago is inadequate for the job and that they need to move the server to another cabinet (they haven't long put it into this one). Anyway, enough rambling, my question is this:-Given that the power failed suddenly, what should I do to confirm that the SQL server and in particular one of the databases, is ok now that it is back up and running?I ask this way as I would like to confirm that what I have done is complete and comprehensive.Many thanks in advancesteveTo alcohol ! The cause of - and solution to - all of life's problems |
|
|
Kristen
Test
22859 Posts |
Posted - 2004-11-01 : 16:45:56
|
| An intermittent fault on a disk controller could be pretty bad news for SQL Server 'coz SQL assumes that when it says "Write to disk" and then disk controller says "I've done that" that it has actually happened!OTOH, for an otherwise healthy system, when the power goes off, and comes back on, SQL just goes "Oh s..." and rolls forward any complete transactions, and throws away the unfinished ones. So no drama.I reckon you've got two options:One is to run a health check with DBCC CHECKDB (See http://www.sqlteam.com/forums/topic.asp?TOPIC_ID=41617)The other is to tell your boss that the IT department's incompetance has caused major damage to the server because of heat-sink, and the server's components will have been heat-stressed such as to make them unreliable in future. The server will be OK for an in-department non-critcal developement-server, but not for 24/7 critical usage. Therefore they should buy a new server immediately (perferably with much faster processors because they will complete the tasks more quickly, and therefore generate much less heat thus preventing a reoccurence) and they should ship the charred, flame-grilled, server over to your department immediately. Or to your home.I haven't had you CV yet, some problem?!Kristen |
 |
|
|
elwoos
Master Smack Fu Yak Hacker
2052 Posts |
Posted - 2004-11-01 : 18:11:27
|
| Thanks Kristen, much appreciated.The DBCC CHECKDB command is what I had in mind but was wondering if I should do anything else really, new server may be a possibility, I would like to upgrade to 2000 anyway so this may be a good excuse, only problem is the 2000 (currently not very critical) server I have went down this morning, I presume for the same reason (the network guys are far to busy to explain what they have resolved or what they plan to do to prevent a repeat, at least for about a week or two until I happen to bump into them).Might look into running the system from my laptop ;)As for my CV, it's well and truly up to date (in fact am currently hoping for an interview on monday). Somehow I doubt you'd be impressedsteveTo alcohol ! The cause of - and solution to - all of life's problems |
|
|
|
elwoos
Master Smack Fu Yak Hacker
2052 Posts |
Posted - 2004-11-02 : 09:07:16
|
| Update that to HAVE an interview on monday :))steveTo alcohol ! The cause of - and solution to - all of life's problems |
|
|
|
Kristen
Test
22859 Posts |
Posted - 2004-11-02 : 09:11:14
|
| What time shall I expect you? <vbg>Kristen |
|
|
|
jen
Master Smack Fu Yak Hacker
4110 Posts |
Posted - 2004-11-02 : 22:35:56
|
quote:
Originally posted by Kristen
What time shall I expect you? <vbg>Kristen
you're scaring him kristen... --------------------keeping it simple... |
|
|
|
Kristen
Test
22859 Posts |
Posted - 2004-11-03 : 01:28:16
|
| Hehehehe!Kristen |
|
|
|
|
|
|
|
|
What now!!