Large Forum -> 2.0 -- step 12 has halted!

[Overgrow]

I'm trying out an upgrade to 2.0.1 on my test box. Made it to step 12 where it has been sitting for 53 minutes. The DB I'm using has about 180,000 posts and I see that it is trying to re-index them.

This test box is a wimpy RAQ4i but an hour ?? Is this normal for step 12 ?

[klisis]

Please copy and paste the actual database error message by viewing source code.

[Mike Sullivan]

It shouldn't take THAT long. Last time I did an upgrade for someone, it took 2 minutes or so.

Although it wasn't a RAQ.

I'd suggest checking what keys exist on the table -- it's possible MySQL got itself all confused (I had someone else's do that back in beta 4 -- it just had to be restarted).

[Overgrow]

hehe no laughing at my testRAQ..

Thanks klisis, there is no error message.. just sitting there on step 12 without progressing.

Ed: must be something I've done. I indexed my tables myself so it must have something confused. I'm going to reimport the DB, remove all indices from post and try again.

thanks~ check back in 20 mins to see if I'm better

[Overgrow]

Hmmm I can't seem to drop the indices on my post table. Could it be because I have two indices named the same thing?

Right now I'm trying to drop them with phpmyadmin (or the command line, whatever) and it hangs here as well. The process list shows that it is "copying to tmp table" but it still sits.

Any help dropping these indices? This is still holding up step 12. I'm being pretty brave trying this upgrade and it hasn't been too bad yet, but its starting...


Here is my post table:

Keyname - Unique - Field

PRIMARY - Yes - postid

idxdisp - No - threadid

idxdisp - No - dateline

[Mike Sullivan]

Originally posted by Overgrow
Hmmm I can't seem to drop the indices on my post table. Could it be because I have two indices named the same thing?

You actually only have one idxdisp index, it's just a multiple column index (see my SitePoint article for more info ).

I'd recommend running the table through myisamchk for index corruption to see if that helps. If it doesn't, we'll just make a post2 table, but I won't explain that unless myisamchk doesn't do anything.

[Overgrow]

myisamchk went smoothly, no errors reported.

Shall we now create a new post table without the indices? Step-by-step appreciated as always Ed!


[root overgrow]# myisamchk post
Checking MyISAM file: post
Data records: 152218 Deleted blocks: 0
- check file-size
- check key delete-chain
- check record delete-chain
- check index reference
- check data record references index: 1
- check data record references index: 2
- check record links

[Mike Sullivan]

Try this:

CREATE TABLE post2 (
postid INT UNSIGNED NOT NULL AUTO_INCREMENT,
threadid INT UNSIGNED NOT NULL,
username CHAR(50) NOT NULL,
userid INT UNSIGNED NOT NULL,
title CHAR(100) NOT NULL,
dateline INT NOT NULL,
pagetext MEDIUMTEXT,
allowsmilie SMALLINT NOT NULL,
email SMALLINT NOT NULL,
signature SMALLINT NOT NULL,
ipaddress CHAR(16) NOT NULL,
iconid SMALLINT UNSIGNED NOT NULL,
visible SMALLINT NOT NULL,
PRIMARY KEY(postid)
) SELECT postid, threadid, username, userid, title, dateline, pagetext, allowsmilie, email, signature, ipaddress, iconid, visible FROM post;

If that doesn't work, we need to do it in 2 steps:

CREATE TABLE post2 (
postid INT UNSIGNED NOT NULL AUTO_INCREMENT,
threadid INT UNSIGNED NOT NULL,
username CHAR(50) NOT NULL,
userid INT UNSIGNED NOT NULL,
title CHAR(100) NOT NULL,
dateline INT NOT NULL,
pagetext MEDIUMTEXT,
allowsmilie SMALLINT NOT NULL,
email SMALLINT NOT NULL,
signature SMALLINT NOT NULL,
ipaddress CHAR(16) NOT NULL,
iconid SMALLINT UNSIGNED NOT NULL,
visible SMALLINT NOT NULL,
PRIMARY KEY(postid)
);
INSERT INTO post2 SELECT postid, threadid, username, userid, title, dateline, pagetext, allowsmilie, email, signature, ipaddress, iconid, visible FROM post;

[Overgrow]

First one wouldn't work.. second one generates:

Got error 28 from table handler

Which probably means "stop trying to run such a big query on such a lame box (256 mb ram? hah)"

[Overgrow]

OK I wrote a small script to move over the data from post to post2. I can do it in batches, but no matter how small the increments, it always halts at around 22,000 records in post2.

Myisamchk reported no errors.........

next step?

[George L]

Originally posted by Overgrow
First one wouldn't work.. second one generates:

Got error 28 from table handler

Which probably means "stop trying to run such a big query on such a lame box (256 mb ram? hah)"

that error means no space left on device ... you sure your hdd or partition has enough free space ?

[Overgrow]

Thanks Eva, you're exactly right. I got home to find the tech-emails waiting saying that I was out of HD space. The POS Raq can't handle any processes modifying such a large DB.. it's just as it comes from the factory right now.