3 Replies Latest reply on Jun 2, 2004 12:18 PM by Jorge Fernandez

    Sybase is Down - Need some help

    Jorge Fernandez
      Share This:

      An old Command/Post - Ensign Console stopped processing alerts, while I'm migra-
      ting to a new PEM - Patrol Console.

      At the sybase errorlog I found the following:
      00:2004/05/11 16:42:34.56 server Recovering database 'agc'.
      00:2004/05/11 16:42:34.57 server Recovery dbid 5 ckpt (460787,24) oldest tran=(460787,23)
      00:2004/05/11 16:42:34.84 server 9 transactions rolled forward.
      00:2004/05/11 16:42:34.84 server 2 transactions rolled back.
      00:2004/05/11 16:42:34.90 server Error: 1105, Severity: 17, State: 3
      00:2004/05/11 16:42:34.90 server Can't allocate space for object 'syslogs' in database 'agc' because the 'logsegment' segment is full. If you ran out of space
      in syslogs, dump the transaction log. Otherwise, use ALTER DATABASE or sp_extendsegment to increase the size of the segment.

      Has somebody in the list, faced this problem??

      The last database backup failed with the same error

      Thanks in advance
      Jorge





      I found some additional info with Sybase stored procedures:
      output from sp_helpdb
      name db_size owner dbid created status
      -


      -


      -


      -


      -


      -



      agc 900.0 MB NetCmmnd 5 Dec 26, 1999 select into/bulkcopy, trunc log on chkpt

      master 5.0 MB sa 1 Jan 01, 1900 no options set

      model 2.0 MB sa 3 Jan 01, 1900 no options set

      sybsystemprocs 64.0 MB sa 4 Dec 26, 1999 trunc log on chkpt

      tempdb 30.0 MB sa 2 May 11, 2004 select into/bulkcopy


      output from sp_helpdb agc
      name db_size owner dbid created status
      -


      -


      -


      -


      -


      -



      agc 900.0 MB NetCmmnd 5 Dec 26, 1999 select into/bulkcopy, trunc log on chkpt


      device_fragments size usage free kbytes
      -


      -


      -


      -



      master 150.0 MB data and log 0
      master 300.0 MB data and log 0
      master 450.0 MB data and log 0


      So, log and data segments are on the same device, with 2 extends up to 900 MB. No space free for log, neither for data

      But, there is enough space on disk to enlarge the database up to 2GB. Sybase re-
      sides on a raw device, at a SUN workstation.


      Output from sp_helpdevice
      device_name physical_name description status cntrltype device_number low high
      -


          -


        -


        -


      -


      -


      ---    -



      agcdump /dev/rmt/0mn tape, 150 MB, dump device 16 3 0 0 4800

      master d_master special, default disk, physical disk, 2000.00 MB 3 0 0 0 1023999

      sysprocsdev /usr/sybase/sysprocs.dat special, physical disk, 64.00 MB 2 0 1 16777216 16809983


      This output confirms what I see from the OS: there is about 1.1 GB of free
      space in master device, to enlarge de DB.


      In this situation, there is no other way to recover than execute a dump of the
      transaction log??
      If so, taking into account that some transactions weren't committed,
      what about consistency of data, in the data segment?? In other words, after
      a dump transaction log, with missed transactions, would the 'agc' database
      recover, and data will be available??

      Again, thanks in advance

        • 1.

          Jorge,

          Not too sure about the data integrity part but it I think this is a simple case of your log running out of space. what you need to do is contact support and they will tell u what to do (I think this will be quicker as I have the notes somewhere but not to hand !!)

          then I would create an extra device to extend the log segment onto...

          again support can send you a step by step guide....on how to do this.....

          I will have a look for the notes anyway....


          Mike.

          • 2.

            Jorge,

            the following resolution id

            138951.......

            Detail....


            Symptom/Question
            Error " DataServer message 1105, state 1,severity 17 "

            Additional Problem or Contact Information
            The error message dbi: DataServer message 1105,
            Can have a number of meanings and understanding these meanings will help you take the right course of action.

            Resolution

            The error DataServer message 1105, state 1,severity 17: Can't allocate space for object 'events' in database 'agc' because the 'default' segment is full refers to the AGC database
            or
            The error DataServer message 1105, state 3,severity 17: Can't allocate space for object 'syslogs' in database 'agc' because the 'logsegment' segment is full refers to the AGC transaction Log

            So if the state = 3 and object = syslogs then it is the transaction log segment of the agc database that is full and can be dumped by doing the following. Note, you can only dump data from the transaction log that has already been written to the data segment.

            isql
            1> dump transaction agc with no_log
            2> go

            So if the state = 1 and object = events or alerts then it is the data segment rather then the log segment that is full. The first step would be to truncate the transaction log as this would free up space if the DATA and LOG parts of the database were on the same segment.

            isql
            1> dump transaction agc with no_log
            2> go

            Then we need to see how much space is available

            isql
            1> sp_helpdevice
            2> go

            isql
            1> sp_helpdb agc
            2> go

            Compare the size of the actual devices compared to the size of the database or log segment located on them and this will show you how much space is available (if any) to expand either the data or log segments.

            If you don't have any space to expand then you then have two options:-

            1) reduce the amount of data in the DB, by reducing the retention period and running doDaily/handleDaily to clear out closed alerts/events

            2) Increase the size of the database by creating a new device and altering the data and/or log segments to use the new device.


            then if I was you I would definatley create another device and then extend the log onto this.

            Hope this helps...


            Mike

            • 3.
              Jorge Fernandez

              Thanks for the help

              Jorge