Search results

To find the root cause ones need: 1. to gather as much information as possible; 2. to have a qualified support team who is able to analyze the huge volume of the collected information; 3. to be lucky to get the answers after the minimal number of the incidents or to be patient to continue an...

All 5 old WS agents (PIDs 14181,14184,14192,14195,14201) died with memory violation between 08:23:21 and 08:23:38 All 5 new WS agents (PIDs 31660,31673,31687,31697,31707) were LOCKED. The reason must be common to them. > also during this time java update was going on. Just curious what the...

Oops! That is why I don't like the numbers! My friend the promon confirms: 1-Single, 2-Stack, 3-One Time

Disagree is my second name. :cool: CachingType = 1 methord: You set the value only once and check the statement cache serially - at t1, t2, t3 etc time. Let’s, every minute. You will get the call stacks on all borders of stat intervals. CachingType = 3 methord: I set CachingType = 3 at t1...

@TomBascom You know the customer of ours who taught me to be very careful about the code that is proposed to run in the production environment. :-) Does anyone have a code to monitor the long running transactions they can share? Long transaction watchdog? I know there is a simple code in Dan...

BTW, if a process rolls back the long transaction then the changes of _UserLock, _UserTableStat and _Trans are the only useful information we can get. proGetStack will not report a call stack.

How often the client statement cache is updated? Once per db request? How many db requests per second does an active client generate? The hundred times? The thousands? Each update of the statement cache creates an extra USR latch lock. Each cache update /may/ create an extra disk io - if a call...

I don't have a ready-to-run code example but the idea is simple: /* ipConnectUsr is a user number */ FOR FIRST DICTDB._Connect NO-LOCK WHERE DICTDB._Connect._Connect-Id EQ ipConnectUsr + 1 TRANSACTION: FIND CURRENT DICTDB._Connect EXCLUSIVE-LOCK. ASSIGN...

BTW, you can't get the program names while a transaction is in an undo phase. Neither with _Connect-Cache* nor with kill -SIGUSR1. I guess the same is true for a subtransaction undo but I did not check it.

You can use: _Connect._Connect-Cache* to get the program name run by the user. _UserLock._UserLock-Table - inte[512] -> Table # _UserTableStat with the increments of _UserTableStat-update or _UserTableStat-create -> _UserTableStat-Num = Table # Since 11.7: _Trans: no increments of...

_Area._Area-attrib is a bitmap and only a few of its 32 bits seemed to be currently used by Progress. Development could dedicate another 3 bits to set the prohibitions for the object types per area level. And Data Dictionary could check these bits. Dreams... We agree - each area should store...

Holy war started! ;-) Arguments to use RPB 1 for the index areas: 1) It's the only way to mark an area to be used for indexes only. We can write a program that will issue a warning if it finds the objects of the other types than the indexes; 2) If someone puts a table in area with RPB 1 and the...

What is the maximum size that can be expected from a field named “ship-city”? For example, the alternate name of Bangkok is 168 character long and it’s mentioned in Guinness Book of Records. Field size statistics from a database somewhere in America ;-) Size Cnt 2332 7 1058 1 1053 1 1051 1...

Or a presentation for then next Progress Conference! I’m glad that now there is a modern guidance about object assignment that we can refer to. > LOBs are inherently large, so it doesn't make sense to me to have an area for miscellaneous LOBs 1. LOBs are the records whose minimal size is ten...

I missed this point. Thanks! > if record sizes are quite uniform LOBs are quite opposite!

The bkLastBiNote field is not used, but is there any rumor as to what it was intended for? Block (the bunch of records) vs an address of some recovery note?

May I ask? If we would be able to change the space management algorithms... When new fragment is added to a first block on RM chain and now the free space of the block is less than the toss limit, would you remove the block from the chain immediately or let the next record create to do that? I...

RMBLK: 0040 numdir: 0x02 2 0041 freedir: 0xfd 253 0042 free: 0x1f8c 8076 0044 entry # 0: 0x1fef 8175 0046 entry # 1: 0x1fd4 8148 "numdir" and "freedir" are 1 byte fixed fields. "free" is 2 bytes. Offset...

You are correct: free space in data block is decreased by RECORD-LENGTH(new record fragment) + 4 bytes: 2 bytes in a record offset directory and another 2 bytes store the size of record fragment. I'm tempting to call these 2 bytes of fragment's size plus the fragment's body itself as a "record"...

Grumbling about Create Limit is continued... I know for sure that Progress technical support has an internal utility that scans AI files and reports the sizes of the recovery notes. The variable part of the sizes of RL_RMCHG notes is directly related to the change the record’s sizes. Progress...