Error in Event Viewer

[Jacky]

Hello,

I still have the following error messages (many many many) in my Event Viewer :

D:\PROGRA~1\SOE\mysql\bin\mysqld-nt.exe: Can't open file: 'history_102006.MYI' (errno: 145)

For more information, see Help and Support Center at http://www.mysql.com.

Source : MySQL
ID event : 100

For weeks, I received from Somix a command to repair :

SOE\mysql\bin>mysql -u root --password=soe -e "repair table perfrpt.history_102006 use_frm;"

It works for a moment but the problem comes ever again. Any Idea ?

Thanks for Help

[tomp]

Hi Jacky,

The command you mentioned does repair the table and will resolve your problem. What we need to investigate is why your history table is becoming corrupt.

Do you have active virus scanning enabled or a backup application that runs each night? If you answered yes to either of these questions, you NEED (and I can't stress this enough) to make sure you have the follow directory exclude: \SOE\mysql\data\

Active (on access) virus scanning software and backup applications tend to lock files when performing their tasks. If a database table is being written to and the file becomes locked it will corrupt the table.

HTH

[Del]

Jacky

I had this same problem. I had to do the repiars almost daily. Jame's helped me out by doing the following.

First check your error log file. C:\SOE\mysql\data\servername.err

My error log was huge. It had the following error:

061030 10:30:48 InnoDB: Started; log sequence number 0 43644
C:\SOE\mysql\bin\mysqld-nt.exe: ready for connections.
Version: '4.1.15-nt' socket: '' port: 3306 Official MySQL binary
061030 10:30:59 [ERROR] C:\SOE\mysql\bin\mysqld-nt.exe: Can't open file: 'network_adapter.MYI' (errno: 145)


We looked through the tables to find 'network_adaptor.MYI' did a repair on it and the error log has yet to append more data. I hope the problem is fixed.

Del

[sbarbee]

I'm receiving a similar error message to the post above, but the suggested repair table doesn't work for me. Below is my error message.

Thanks,
Scott

Event Type: Error
Event Source: MySQL
Event Category: None
Event ID: 100
Date: 1/28/2008
Time: 9:05:19 AM
User: N/A
Computer: ***REMOVED***
Description:
C:\SOE\mysql\bin\mysqld-nt.exe: Can't open file: 'history_012008.MYI' (errno: 145)

For more information, see Help and Support Center at http://www.mysql.com.

[tomp]

Open a command prompt, navigate to C:\SOE\mysql\bin\ and run the following command:


mysql -u root --password=soe -e "repair table perfrpt.history_012008 USE_FRM\G;"

Let me know if you still have problems after running this repair.

[sbarbee]

Tomp,

The command you provided seems to have worked. No error logs in 15 minutes, where previously I was getting around 1000/min before.

Thanks,
Scott

[tomp]

Database tables can become corrupted if an active virus scanning software or backup application (ie: Backup Exec) attempts to hold one of the database files (ie: history_012008.MYI)

It is recommended to exclude \SOE\mysql\data\ from active virus scanning and backup applications. As an alternative, perform a scheduled virus scan and use MySQL Administrator to backup the database to another directory that is being backed up by your backup application.

[isd340]

I just moved SOE to a different drive on the same server (same IP). This is a sampling of the errors I'm getting:

logalot.eventlog_history_122008: Table upgrade required. Please do "REPAIR TABLE `eventlog_history_122008`" to fix it!
logalot.history_122008: Table upgrade required. Please do "REPAIR TABLE `history_122008`" to fix it!
logalot.orphan: Table upgrade required. Please do "REPAIR TABLE `orphan`" to fix it!
perfrpt.vendorindex: Table upgrade required. Please do "REPAIR TABLE `vendorindex`" to fix it!
security.useraccounts: Table upgrade required. Please do "REPAIR TABLE `useraccounts`" to fix it!
toc.navigation: Table upgrade required. Please do "REPAIR TABLE `navigation`" to fix it!

[tomp]

The errors you are seeing are related to the MySQL version. This isn't a big problem. I've seen with MySQL upgrades and migrating old SOE installations to new SOE installations.

There are a couple different ways to fix this. Each will involve having to repair the tables just as the error message says. You can either do it command line or with a MySQL gui.

If you want to do this via command line, you can use the following command

mysql -u root --password=soe -e "repair table DATABASE.TABLE USE_FRM\G;"

Open a command prompt on the server and navigate to the \SOE\mysql\bin directory and run the command. I have created all of the commands for the samples you have listed:

mysql -u root --password=soe -e "repair table logalot.eventlog_history_122008 USE_FRM\G;"

mysql -u root --password=soe -e "repair table logalot.history_122008 USE_FRM\G;"

mysql -u root --password=soe -e "repair table logalot.orphan USE_FRM\G;"

mysql -u root --password=soe -e "repair table perfrpt.vendorindex USE_FRM\G;"

mysql -u root --password=soe -e "repair table security.useraccounts USE_FRM\G;"

mysql -u root --password=soe -e "repair table to.navigation USE_FRM\G;"

The other option is to use MySQL Administrator, which is a great MySQL GUI. You can download it from: www.mysql.com. Once you download and install. Launch the application, connect to the database using the root account. The password is defaulted to: soe

Once you are in the application click on "Catalogs" on the left hand side and then click on one of the affected databases. It will load all of the tables in the right-hand pane. Select all of the tables and click the "Maintenance" button. Click "Repair Tables", then select "Quick" for the Repair Method. It should be the only option you need to select. Finally, click "Repair Tables". It should come back with the list of tables with a status of OK.

Click on the next affected database under "Catalog". You will need to repeat for the all databases exhibiting this problem. The repair process shouldn't take too long. Some of the Logalot history tables may take a little time if they are excessive in size.

I'm curious. What process did you use to migrate to another drive? Did you make a backup, uninstall the application, then re-install using a new install shield?

Let me know if you have any problems.

[isd340]

Thanks for the great instructions. I did a few with the command line but decided to try the MySQL Admin tools since I'm a Network admin want to be Sys admin I was able to repair all the databases except 'information_schema' probably because it's live? Do I need to stop some services and try to repair it? I still do not have any graphs in my reports.. Do I need to stop some service and do the repair on the schema?

To answer your question on my method of moving to another drive, I found a useful set of instructions from a post on this site entitled Post subject: Move Denika 6.5 to another server.. In another post I read that I could just copy the SOE folder to the new location - that's what I did after installing the more recent version of Denika ( The different version numbers is probably where I went wrong).

Another issue I'm having is that I see the Evaluation timer ticking on the Admin page ("Evaluation Time Left: 29 days."). Under Denika Config > General Licensing I see license numbers for Denika and Logalot - I will look for our actual license numbers and compare them to make sure the ones listed are not just temporary license numbers.

If you can help me find out why I'm not showing any report graphs that would be great. If I have to scrap everything and start over -that's ok as well. It's just that one of our devices, a Cisco 6509, had to have custom mibs defined in order to coordinate the right graph with the right interface instance.

[tomp]

The information schema is a built-in mysql table. There is no need to repair that table.

Now, for viewing your reports. I think because you changed the drive, you may need to change the drive letter in the SOE configuration. Launch the web interface, navigate to MENU -> Control Center -> Configuration Editor and click on "General" under SOE and make sure you are using the correct Drive letter for SOE Install Path.

[isd340]

Thanks, I had already done that with no change.. I went through every instance that I could find in the registry and changed all C: references to D: - still nothing..

On the good side, I stumbled across another almost unrelated post that talked about not using the ff browser (which I almost exclusively use). I tried a discovery with ie and now I can make new graphs! Still don't have my 100 + old ones but at least there's some light at the end of the tunnel and I can create new reports to address immediate network questions.

[jghidoni]

Have you already confirmed that your reports are not all disabled in Denika?

On the main Report page in Denika, bottom right of main window, there is a link that says "Disabled Reports". does the number of Disabled Reports match that of "Reports" just to the left of that?

Or, can you verify that the Cron service is running, which runs the poller process? If it is, can you check the timestamps of the pollFile_x.* files in soe\mrtg\bin? Are the timestamps within the last 5 minutes, or are they old? Specifically looking for pollFile_x.ok files, and that they are within the last 5 minutes, and that there are no pollFile_x.cfg_l files with old timestamps.

If there are cfg_l files with old timestamps, possibly the poller.exe and mrtg.exe processes have got hung up. If that is the case, kill the poller.exe and mrtg.exe processes, then delete any soe\mrtg\bin\pollFile_x.* files, and run soe\cgi-bin\poller.exe from command prompt and see if any errors, or wait 10-15 minutes to see if reports collecting data now.

- Joanne

[isd340]

Hmm, you caught me on one front.. All my reports were indeed disabled However enabling them did not give me any graphs.

Cron service has been running all this time.

can you check the timestamps of the pollFile_x.* files in soe\mrtg\bin?

The pollFile_1.ok file is 0K and doesn't have anything in it. The pollFile_1.cfg has the new reports I just discovered today but nothing else.

Specifically looking for pollFile_x.ok files, and that they are within the last 5 minutes, and that there are no pollFile_x.cfg_l files with old timestamps.

Since I enabled all the reports, I now have a pollFile_1.cfg_| file. Inside it says: 3936

kill the poller.exe and mrtg.exe processes, then delete any soe\mrtg\bin\pollFile_x.* files, and run soe\cgi-bin\poller.exe from command prompt and see if any errors, or wait 10-15 minutes to see if reports collecting data now.

I don't see a poller.exe or an mrtg.exe process running in services, do you want me to just stop the cron service?

[jghidoni]

Hi,

Okay, this doesn't look too bad then. Don't stop the Cron service, but do delete the pollFile_1.* files in soe\mrtg\bin, then run soe\cgi> poller.exe from command prompt, and it will recreate the pollFile_1.cfg and possibly a pollFile_2 if you have enough reports.

Then after 15-20 minutes, you should start seeing data in the old reports.

- Joanne