Home
Jason L. Froebe
Me :)
Recent Entries 
blue man

There are many former Sybase employees that retain the knowledge of the Sybase ASE internals.  No, not the internals from the Data Internals class but the source code and design documents.  While we gave up the physical materials when we left Sybase, we still retain much of it in our brains.  Technically, we aren’t supposed to use that knowledge but it is very very difficult not to.

Think about this scenerio:

Corruption on a single page in a table.  Sounds easy, just bcp the data out, truncate the table, bcp the data back in.  Normally that is the way to fix it.

What if the system is a mission critical 24×7x365 system that might be used for international trading or even to provide medical information to medical surgeons?  BCPing the data or going to backups may not be possible.  If you had the information on how to fix the corrupted page by patching it (something that Sybase strongly frowns upon)?

As you can see, using the internal/proprietary information inside your head is hard to avoid.  Granted, ProActive DBA allows you to patch the individual pages, but they apparently have permission to do so — or at least Sybase is looking the other way.

If you know about SybMon and know how to use it, it is nearly impossible not to use it if your Sybase ASE server is hung and you can’t log into it.

If you fix a server using an unsupported method and report it to Sybase, there is a chance that they will not look favorably on your solution.  So, do you report the problem to them and your solution in the hope that they fix the problem or do you just keep it to yourself?

Harder yet, is how to explain to your boss that you can’t tell them how you fixed the problem because you’re still under a non-disclosure agreement with Sybase.  You might be able to explain the situation to your boss but if he pushes for the information and you refuse, you might be out of a job.  Luckily my boss is happy to have me and doesn’t push the issue. :)

blue man
Does anyone know the syntax for the dbcc dbrepair commands writepage and readpage for Sybase ASE?  These are being used by Whitesands's ProActive DBA.

The equivalent on Microsoft SQL Server are:
dbcc readpage ({ dbid, 'dbname' }, fileid, pageid, formatstr [, printopt = { 0 | 1} ])
dbcc writepage ({ dbid, 'dbname' }, fileid, pageid, offset, length, data)

Read more... )

13th-Jul-2006 02:39 am - work...
blue man
Been working 15 to 18 hour days this week and today a 10 hour day. We are in a period of adjustment where we fix the environment (database-wise) across Solaris, AIX, Win2k, and Linux servers in various locations. Monitoring scripts for performance, backup verifications, configuration, etc. are in the process of being written.

Politics all around as the previous dba was stretched way too thin to be very proactive (it wasn't his fault) and certain individuals look upon our team as country bumpkins. Meaning unreliable. It isn't against the individuals of our group but just the idea of our group's existance. I think they don't like having to defer to our group for administration of the databases. Boy are they in for a surprise once we get the environment up and running... security crack down.

You know what? DB Artisan is the biggest POS I've run into for a DBA front end since Sybase's Sybase Central. It is bloated, doesn't recognize when a connection disconnects, allows only one active working connection (ie: run a query), and unstable. Yes, it is the very latest version. Guess what? It doesn't support Sybase's Open Client 15 so it is of limited use. POS!

Apparently v12.5.3 esd 7 64bit on Linux has a problem with the backup server. It throws a signal 11 and dies. Nice feature, huh? The problem is that esd 7 introduced a 64bit backupserver binary and still uses the old compat-libcstd++-5 library. The 64bit version of the library is somewhat flakey. Using the 32bit binary from esd 6 is the workaround (make sure to include the esd7 version of sybmultbuf too). I'm waiting on the 12.5.4 from Sybase (I get the cds :), so hopefully they fixed it.
blue man
http://www.sybase.com/detail?id=1040311 The techdoc describes that data may be lost due to os buffering of the block devices (raw devices) if the machine goes down.  Unfortunately, it doesn't indicate whether "shutdown with nowait", "kill -9", "kill -15" or "kill -4" also result in data loss because of the os buffering.

I asked Sybase techsupport to update the techdoc and the cr.

PSE has updated the CR 406473 description )
blue man
I've deprecated the sybase-dba@yahoogroups.com mailing list in favor of sybase-dba-moderated@yahoogroups.com mailing list due to the spamming problem on sybase-dba.

A moderated discussion of administration of Sybase databases, Sybase Replication, etc.

Databases covered:
Adaptive Server Enterprise (ASE)
SQL Anywhere (AKA Adaptive Server Anywhere)
Adaptive Server IQ (ASIQ)

Yahoo Group: http://groups.yahoo.com/group/sybase-dba-moderated/
Post message: sybase-dba-moderated@yahoogroups.com
Subscribe: sybase-dba-moderated-subscribe@yahoogroups.com


Subscribe to sybase-dba-moderated





Powered by groups.yahoo.com

blue man
select count(1) "Total Locks" from syslocks

select t.spid, db_name(syslocks.dbid) "Database", object_name(syslocks.id, syslocks.dbid) "Object Name", count(t.loid) "Total Locks"
from systransactions t, syslocks
where t.loid = syslocks.loid
group by syslocks.dbid, syslocks.id

Yes, it is easier to use the MDA tables in Sybase ASE (12.5.0.3 and higher) but there is a security risk with the MDA tables that is currently implemented:

Anyone with the mon_role has the ability to see other spids' sql text. Not bad? Well, when you change your password (sp_password) the old and the new passwords are viewable! Supposedly Sybase will fix this in v15.1.
This page was loaded Dec 3rd 2009, 12:50 am GMT.