In Part I of this blog post I covered the specifics of the -519 Jet_errLogSequenceEnd issue and the resolution. We had a great follow up question from our readers about the when and how often Exchange issues the ESE 514 warning events? We'll answer those questions in detail here in Part II.
How much time do I have?
The key information you should focus on in the ESE 514 event is the current log generation. Have a look at our TechNet document that explains this event (you'll notice this is for Exchange version 8.0 aka. Exchange 2007; it functions the same way for Exchange 6.5 and 6.0 or Exchange 2003 and 2000 respectively): http://www.microsoft.com/technet/support/ee/transform.aspx?ProdName=Exchange&ProdVer=8.0&EvtID=514&E...
Here is the event string with variables:
|
Product:
|
Exchange
|
|
ID:
|
514
|
|
Source:
|
ESE
|
|
Version:
|
8.0
|
|
Symbolic Name:
|
ALMOST_OUT_OF_LOG_SEQUENCE_ID
|
|
Message:
|
%1 (%2) %3Log sequence numbers for this instance have almost been completely consumed. The current log generation is %4 which is approaching the maximum log generation of %5, there are %6 log generations left to be used. To begin renumbering from generation 1, the instance must be shutdown cleanly and all log files must be deleted. Backups will be invalidated.
|
Here is what this looked like on an Exchange 2003 server in production (from a support case I recently worked on):
Event ID : 514
Category : None
Source : ESE
Type : Warning
Machine : <Exchange Server Name>
Message : Information Store (5144) <Store Name>: Log sequence numbers for this instance have almost been completely consumed. The current log generation is 1044490 (0x000FF00A) which is approaching the maximum log generation of 1048559 (0x000FFFEF), there are 4069 (0x00000FE5) log generations left to be used. To begin renumbering from generation 1, the instance must be shutdown cleanly and all log files must be deleted. Backups will be invalidated.
Unfortunately these warnings were missed, and the server eventually reached:
Event ID : 514
Category : None
Source : ESE
Type : Warning
Machine : <Exchange Server Name>
Message : Information Store (5144) <Store Name>: Log sequence numbers for this instance have almost been completely consumed. The current log generation is 1048550 (0x000FFFE6) which is approaching the maximum log generation of 1048559 (0x000FFFEF), there are 9 (0x00000009) log generations left to be used. To begin renumbering from generation 1, the instance must be shutdown cleanly and all log files must be deleted. Backups will be invalidated.
All the ESE 514 Events were missed and Exchange shut down the databases not long after the second event was logged. Recall from Part I of the blog that Exchange issues the MSExchangeIS 1159 error when transaction log sequence numbering has been exhausted.
What's interesting about these events is that Exchange tells you exactly how long you have before all log file generations are used up and we increase the frequency of warning messages the closer we get to Jet_errLogSequenceEnd.
When do the warnings start and how often do they occur?
For Exchange 2000, 2003, and 2007, we use thresholds to determine when and how often the ESE 514 warning events will be fired. Those thresholds are different for Exchange 2000/2003 and Exchange 2007 due to Exchange 2007's much longer transaction log sequence range. Recall from Part I of the blog the maximum transaction log sequence number in Exchange 2000/2003 is FFFF0 (1,048,560 decimal) while Exchange 2007 is 7fffffec (2,147,483,628 decimal).
The following sections describe the exact thresholds that trigger the first ESE 514 warning events and frequency at which subsequent warning events will be fired.
Exchange 2000/2003:
Exchange 2000/2003 will fire the first ESE 514 Event at log generation number 0xE0000 (917,504 decimal) for a given Storage Group. From that point, Exchange fires an ESE 514 warning event every 1000 logs until we hit log generation 0xFF000 (1,044,480 decimal). By this point Exchange is dangerously close to the maximum and goes into Panic Mode, issuing the Warning every 10 transaction logs until we run out.
Exchange 2007:
Exchange 2007 will fire the first ESE 514 Event at log generation number 0x7FFE0000 (2,147,352,576 decimal) for a given Storage Group. From that point, Exchange fires an ESE 514 warning event every 1000 logs until we hit log generation 0x7FFFF000 (2,147,479,552 decimal); at this point Exchange is dangerously close to the maximum and goes into Panic Mode, issuing the Warning every 10 transaction logs until we run out. If you indeed have run out of Exchange 2007 transaction logs - let us know. We'd love to hear how come that many logs are generated!
Conclusion:
The key to avoiding database shutdown due to -519 Jet_errLogSequenceEnd is to monitor your Exchange application logs. When you start to see ESE 514 Warnings make plans to reset transaction log sequence. If you're seeing those very frequently, better act fast.
Corbin Meek
Enterprise Communications Support
