MySQL error 28

[marcopolo]

I understand this error relates to disk space but would like to know whether it is generated by MySQL or vB and if I can determine the actual file path that it relates to in terms of /tmp

Database error in vBulletin 4.1.3:

Invalid SQL:
SELECT node.nodeid AS itemid,
(node.nodeleft = 1) AS isroot, node.nodeid, node.contenttypeid, node.contentid, node.url, node.parentnode, node.styleid, node.userid,
node.layoutid, node.publishdate, node.setpublish, node.issection, parent.permissionsfrom as parentpermissions,
node.permissionsfrom, node.publicpreview, node.showtitle, node.showuser, node.showpreviewonly, node.showall,
node.showupdated, node.showviewcount, node.showpublishdate, node.settingsforboth, node.includechildren, node.editshowchildren,
node.shownav, node.hidden, node.nosearch, node.nodeleft,
info.description, info.title, info.html_title, info.viewcount, info.creationdate, info.workflowdate,
info.workflowstatus, info.workflowcheckedout, info.workflowlevelid, info.associatedthreadid,
user.username, sectionorder.displayorder, thread.replycount, parentinfo.title AS parenttitle

FROM cms_node AS node
INNER JOIN cms_nodeinfo AS info ON info.nodeid = node.nodeid

LEFT JOIN user AS user ON user.userid = node.userid
LEFT JOIN thread AS thread ON thread.threadid = info.associatedthreadid
LEFT JOIN cms_sectionorder AS sectionorder ON sectionorder.sectionid = 1
AND sectionorder.nodeid = node.nodeid
LEFT JOIN cms_node AS parent ON parent.nodeid = node.parentnode
LEFT JOIN cms_nodeinfo AS parentinfo ON parentinfo.nodeid = parent.nodeid
INNER JOIN cms_node AS rootnode
ON rootnode.nodeid = 1 AND (node.nodeleft >= rootnode.nodeleft AND node.nodeleft <= rootnode.noderight) AND node.nodeleft != rootnode.nodeleft AND node.contenttypeid <> 17 AND node.new != 1 AND ( (( (node.permissionsfrom IN (-1)) OR ( node.permissionsfrom in (1) AND (node.parentnode IN (1,378,194,113,426,114,190,231,232,120,121,123,132 ,296,124,382,134,298,315,133,126,463,138 ,137,139,316,302,146,140,353,303,317,127,465,141,3 58,313,321,198,142,314,143,468,145,128,1 49,356,323,324,150,360,151,337,471,148,147,310,325 ,129,156,318,319,157,330,158,160,308,161 ,162,322,159,464,152,153,309,154,155,130,164,163,1 66,165,167,170,470,169,168,131,173,171,1 72,174,179,175,176,177,178,394,401,402,403,400,404 ,398,399,395,461,396,406,414,411,405,407 ,415,413,410,409,412,397,419,424,418,423,420,422,4 21,425,180,244,366,375,369,187,196,197,1 88,363,195,222,223,224,182,185,191,193,204,351) OR node.nodeid = 1) AND
node.setpublish > 0 AND node.publishdate < 1305601398 ))) OR (node.setpublish AND node.publishdate <1305601398 AND node.publicpreview > 0))AND node.hidden = 0 AND node.setpublish = '1' AND node.publishdate <= 1305601398 AND (node.parentnode = 1 OR sectionorder.displayorder > 0 )

ORDER BY CASE WHEN sectionorder.displayorder > 0 THEN sectionorder.displayorder ELSE 9999999 END ASC,
node.publishdate DESC LIMIT 0, 7;

MySQL Error : Got error 28 from storage engine
Error Number : 1030
Request Date : Tuesday, May 17th 2011 @ 04:03:18 AM
Error Date : Tuesday, May 17th 2011 @ 04:03:18 AM
Script : http://<myhost>./index.php
Referrer : http://<myhost>./index.php
IP Address : 31.31.32.54
Username : Unregistered
Classname : vB_Database
MySQL Version :

I'm querying it because there is a global /tmp and per host /tmp and I want to make sure my supplier is looking at the correct one. I can then make a decision to enable the forum again or not, because to date they have advised nothing has ran out of space.

I also need to confirm that if it does relate to a specific folder, if anything has been increased in the meantime and the likelihood of the /tmp temprarilly filling up.

Any help/guidance appreciated

Mark

PS - historical discussion over at vBulletin.com

[snakes1100]

In a terminal, type "df -h", free up some disk space.

To find large files:
find / -size +512000
find / -size +1024000

[marcopolo]

Originally Posted by snakes1100

In a terminal, type "df -h", free up some disk space.

To find large files:
find / -size +512000
find / -size +1024000

That's not what I'm after.

[snakes1100]

Sorry, i skipped the initial part of your post about tmp, was just focusing on the actual error, which is generated from mysql, not vbulletin.

Edit my.cnf or my.ini and set a new tmp path.

tmpdir = /home/tmp or the largest partition you have.
chown that new directory to mysql
restart mysql

Also, verify that your mysql init script is not forcing a tmp dir when starting sql.

[marcopolo]

Thanks for that

Are those high level files that and since it's a shared server I probably won't have the ability to restart MySQL?

Maybe something I can pass onto the hosting company though I'm suspecting the issue is when a lot of transactions are being submitted (it's maybe periodically filling something up), hence filling up the defined root /tmp folder and I'm only receiving the email errors intermittently.

As you can see, nothing here really suggests anything wrong and maybe they need to reboot the server in the meantime

pop up
cpanellogd up
imap up
ftpd up
exim (exim-4.69-28_cpanel_maildir) up
sshd up
cpsrvd up
mysql (5.0.91-community) up
httpd (2.2.15 (Unix)) up
syslogd up
exim-587 up
lfd up
postgresql up
queueprocd up
Server Load 6.15 (8 cpus)
Memory Used 8.4 %
Swap Used 0.18 %
Disk /dev/shcp02data/data01 (/data01) 65 %
Disk /dev/shcp02data/data02 (/data02) 53 %
Disk /dev/shcp02data/data03 (/data03) 56 %
Disk /dev/shcp02data/data04 (/data04) 43 %
Disk /dev/shcp02mail/mail01 (/mail01) 85 %
Disk /dev/shcp02mail/mail02 (/mail02) 75 %
Disk /dev/VolGroup00/var (/var) 57 %
Disk /dev/VolGroup00/tmp (/tmp) 63 %
Disk /dev/VolGroup00/usr (/usr) 77 %
Disk /dev/root (/) 30 %
Disk /dev/VolGroup00/mysql (/var/lib/mysql) 29 %
Disk /dev/sda1 (/boot) 12 %
Disk /dev/VolGroup00/home (/home) 17 %
Disk /dev/VolGroup00/domlogs (/usr/local/apache/domlogs) 49 %

[snakes1100]

Disk /dev/VolGroup00/tmp (/tmp) 63 %

What is the size of tmp?

If tmp is staying filled like that 63% and tmp is only 1gb, your running a repair on a large table like post, could easily fill it up.

[marcopolo]

Not sure of the actual size because I don't know if this refers to the area I cannot access or my /tmp area

Yeah I guess it comes down to what's available and what processes are being performed. Especially when the errors are reported periodically.

Today Disk /dev/VolGroup00/tmp (/tmp) 67 %

[snakes1100]

You stated that the server is shared, im going to assume its not a vps?

If you have ssh access, you can simply type df -h and get the info about partition sizes & usage.

[marcopolo]

It's one of these http://www.zen.co.uk/business/hostin...b-hosting.aspx

Silver package.

[snakes1100]

Yea, with an account of that type, you'd have no control over anything for the most part, this will fall solely on your host to resolve.