Some Problems with Zabbix 1.6

[mlategan]

Hi there...

I loaded 1.6 but are experiencing some problems. I'm using Postgresql 7.4.

1) Image uploads

When I upload gif icons/backgrounds, I have to do it twice. The first time it uploads, and then gives gd errors where the image preview is supposed to be. If I open the image with the gd errors and redo the upload, it takes it the second time.

2) Auto-discovery does not start.

I have no idea why. If I look at the log, it reports no errors, and states "server #22 started [Discoverer. SNMP:YES]" which I assume means that it thinks the discoverer started. I restarted the service several times and once, without any config changes it did start, and worked like a charm. Until I restarted the service again. Now its dead again and wont start up.

3) Postgresql errors in the log:

  5902:20080919:021432 Type of received value [Collector is not started!] is not sutable for [system.cpu.load[,avg15]@eden_files_ads.edendm.co.za] having type [0]
  5902:20080919:021432 Returning NOTSUPPORTED
  5902:20080919:021432 Parameter [system.cpu.load[,avg15]] is not supported by agent on host [eden_files_ads.edendm.co.za] Old status [0]
  5902:20080919:021432 Query::update items set status=3,lastclock=1221783272,nextcheck=1221783872,error='Type of received value [Collector is not started!] is not sutable for [system.cpu.load[,avg15]@eden_files_ads.edendm.co.za] having type [0]' where itemid=22283
  5902:20080919:021432 Query failed:PGRES_FATAL_ERROR:ERROR:  value too long for type character varying(128)

etc
etc

This might be due to 1.4 agents still running on some machines.

Any ideas what I might have misconfigured? The main issue is the auto-discovery.

Thank you for any suggestions.

[SMB]

Hi there...

I loaded 1.6 but are experiencing some problems. I'm using Postgresql 7.4.

I think you'll get better performance out of upgrading to 8.3

1) Image uploads

When I upload gif icons/backgrounds, I have to do it twice. The first time it uploads, and then gives gd errors where the image preview is supposed to be. If I open the image with the gd errors and redo the upload, it takes it the second time.

1.4 did this as well.

2) Auto-discovery does not start.

I have no idea why. If I look at the log, it reports no errors, and states "server #22 started [Discoverer. SNMP:YES]" which I assume means that it thinks the discoverer started. I restarted the service several times and once, without any config changes it did start, and worked like a charm. Until I restarted the service again. Now its dead again and wont start up.

Personally, I've never seen discovery do anything.

3) Postgresql errors in the log:

  5902:20080919:021432 Type of received value [Collector is not started!] is not sutable for [system.cpu.load[,avg15]@eden_files_ads.edendm.co.za] having type [0]
  5902:20080919:021432 Returning NOTSUPPORTED
  5902:20080919:021432 Parameter [system.cpu.load[,avg15]] is not supported by agent on host [eden_files_ads.edendm.co.za] Old status [0]
  5902:20080919:021432 Query::update items set status=3,lastclock=1221783272,nextcheck=1221783872,error='Type of received value [Collector is not started!] is not sutable for [system.cpu.load[,avg15]@eden_files_ads.edendm.co.za] having type [0]' where itemid=22283
  5902:20080919:021432 Query failed:PGRES_FATAL_ERROR:ERROR:  value too long for type character varying(128)

etc
etc

This might be due to 1.4 agents still running on some machines.

It's not, and this issue existed in 1.4 as well. The problem is that there are fields that are defined as too small for the possible data that could be inserted (value too long for type character varying(128)). The easiest fix is to simply extend the size of the column, you might want to increase your logging level to see exactly which columns in what table(s) are failing.

[mlategan]

Thanks for the reply SMB.

Short update: All seems to be working well now. After a reboot the autodiscovery started up, and the hosts started appearing immediately. Maybe there was still some 1.4 stuff running that messed things up.

On your reply:

Postgres 8.3: Thank you for the advice, will do.

Image Uploads: No biggy, its working so I'm not too bothered about it. I don't recall ever getting it with 1.4 on the same system, though.

Postgres Errors: Thank you, I will extend the columns as suggested.

Many thanks to the development team. This release seems great so far. Time to test all the new stuff!

[mlategan]

Another update:

It seems all problems I was experiencing is related to Postgresql. Some more issues came up:

1) Cannot delete anything.
2) Not all monitoring values are monitored. That is, if I connect (say) the Linux template, only certain values are actively monitored. The others just stay empty. This seems to be random. If I disable and then recreate the exact same item, it sometimes works.

and some other issues here and there.

So I reloaded everything with mysql, and now:

1) I only upload images once. It works 100%.
2) Autodiscovery works from the word go, and has been working now for 3 days.
3) Deleting items/triggers etc is not a problem.
4) All values are monitored properly.

One more comment: The manual says that the initial user is "admin" with no password. This is not the case. The initial user is "Admin" and on both postgres and mysql I had to manually run a query to set the password to something (update users set password = md5('somepassword') where userid=1) to get into the system.

Other than that: With mysql, thus far ,no problems at all. With postgres, its a nightmare. Now just to get mysql not to corrupt its page files and consequently render the database useless on power loss, or other irregular failures. Something I had to deal with three times in the last month

[Tenzer]

One more comment: The manual says that the initial user is "admin" with no password. This is not the case. The initial user is "Admin" and on both postgres and mysql I had to manually run a query to set the password to something (update users set password = md5('somepassword') where userid=1) to get into the system.

The default password is "zabbix" as discussed in this thread: http://www.zabbix.com/forum/showthread.php?t=10547

[teferi]

Another update:

It seems all problems I was experiencing is related to Postgresql. Some more issues came up:

1) Cannot delete anything.
2) Not all monitoring values are monitored. That is, if I connect (say) the Linux template, only certain values are actively monitored. The others just stay empty. This seems to be random. If I disable and then recreate the exact same item, it sometimes works.

and some other issues here and there.

So I reloaded everything with mysql, and now:

1) I only upload images once. It works 100%.
2) Autodiscovery works from the word go, and has been working now for 3 days.
3) Deleting items/triggers etc is not a problem.
4) All values are monitored properly.

One more comment: The manual says that the initial user is "admin" with no password. This is not the case. The initial user is "Admin" and on both postgres and mysql I had to manually run a query to set the password to something (update users set password = md5('somepassword') where userid=1) to get into the system.

Other than that: With mysql, thus far ,no problems at all. With postgres, its a nightmare. Now just to get mysql not to corrupt its page files and consequently render the database useless on power loss, or other irregular failures. Something I had to deal with three times in the last month

Any further comment on postgresql issues? Im expiriencing exact same and well any comment on whether it's possible to mix db engines on proxies and on master node?

[teferi]

well only mine are a bit more severe

 66831:20080924:170942 Query::insert into history (clock,itemid,value) values (1222261683,57039,21162014026969207295668090796145235818888835131879202551875485137018811509127612290830686367610235959536043328121683934802731461179707531745324934682409728126379454099366052859610637647138178248453565087199461900288.000000)
 66831:20080924:170942 Query failed:PGRES_FATAL_ERROR:ERROR:  numeric field overflow
DETAIL:  A field with precision 16, scale 4 must round to an absolute value less than 10^12.

 66831:20080924:170942 Query::select num,value_min,value_avg,value_max from trends where itemid=57039 and clock=1222261200
 66831:20080924:170942 Query failed:PGRES_FATAL_ERROR:ERROR:  current transaction is aborted, commands ignored until end of transaction block

 66822:20080924:170942 One child process died. Exiting ...
 66822:20080924:170944 ZABBIX Server stopped. ZABBIX 1.6.
^C

[SMB]

1) Cannot delete anything.

Are any errors thrown? Does the user your connecting to the database with have delete/update privileges to those tables?

 66831:20080924:170942 Query::insert into history (clock,itemid,value) values (1222261683,57039,21162014026969207295668090796145235818888835131879202551875485137018811509127612290830686367610235959536043328121683934802731461179707531745324934682409728126379454099366052859610637647138178248453565087199461900288.000000)
 66831:20080924:170942 Query failed:PGRES_FATAL_ERROR:ERROR:  numeric field overflow
DETAIL:  A field with precision 16, scale 4 must round to an absolute value less than 10^12.

Do the following against the database you're holding the zabbix tables

ALTER TABLE history ALTER "value" TYPE numeric;

[mlategan]

Are any errors thrown?

Unfortunately I dont have the pg installation anymore, but if I recall correctly, it did not throw an onscreen error. It just went back to the screen where I selected the items to be deleted, and they were still there.

Does the user your connecting to the database with have delete/update privileges to those tables?

I am not a postgresql expert, so please correct me if my assumptions about ownership is wrong:

I created the db with createdb -O zabbix zabbix, which (I assumed) made zabbix the owner of the database, and I ran the creation scripts on all tables in the database using the zabbix user, so I believe that the user had all the right privileges.

[Aly]

ALTER TABLE history ALTER "value" TYPE numeric;

You really shouldn't do it!!!

[teferi]

ALTER TABLE history ALTER "value" TYPE numeric;

this is a dirty hack and it shouldn't really be done, because it'll reduce db speed when relating to this table(very often indeed)
Stil I couldn't suggest anything better yet. so until the fix is avaliable I'll stay with value with arbitrary prescision/scale.

[NOB]

Hi

AFAIR I had to set the value to something like 19,4 to solve this problem.
17,4 didn't work.
Even after looking at the values and their meaning in both MySQL and PostgreSQL, I don't understand why this shouldn't work well with both
DB-engines. For me it's a mystery why I had to add more digits to the
number.

HTH,

Norbert.

[teferi]

I'd suggest adding some checking about whether the value will fit in 16,4 if we use PostgreSQL in zabbix_server code.

Adding rules or triggers made little sense. Propably because the error is issued at parsing stage, when postgres realises that the number is abnormal. Or it's something wrong with proxy side - because such numbers should never come through.

[stevender]

But there still is an error when deleting hosts from zabbix running with Postgre.
I think it has something to do with the auto casting in postgre that was disabled in 8.3 ?

If I delete a host and take a look at the detailed error report:

(ERROR: IN types character varying and integer cannot be matched)

    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: IN types character varying and integer cannot be matched[/var/www/html/include/db.inc.php:390]
    * Error in query [SELECT DISTINCT actionid FROM conditions WHERE conditiontype=2 AND (value IN (13338,13339,13340,13341,13342,13343,13344,13345,13346,13347,13348,13349,13350,13351,13352,13353,13354,13355,13356,13357,13358,13359,13360,13361)) ] [ERROR: IN types character varying and integer cannot be matched]
    * pg_fetch_assoc() expects parameter 1 to be resource, boolean given[/var/www/html/include/db.inc.php:520]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:463]
    * Error in query [DELETE FROM conditions WHERE conditiontype=2 AND (value IN (13338,13339,13340,13341,13342,13343,13344,13345,13346,13347,13348,13349,13350,13351,13352,13353,13354,13355,13356,13357,13358,13359,13360,13361)) ] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:390]
    * Error in query [SELECT triggerid, description FROM triggers WHERE (triggerid IN (13338,13339,13340,13341,13342,13343,13344,13345,13346,13347,13348,13349,13350,13351,13352,13353,13354,13355,13356,13357,13358,13359,13360,13361)) ] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_fetch_assoc() expects parameter 1 to be resource, boolean given[/var/www/html/include/db.inc.php:520]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:463]
    * Error in query [DELETE FROM triggers WHERE (triggerid IN (13338,13339,13340,13341,13342,13343,13344,13345,13346,13347,13348,13349,13350,13351,13352,13353,13354,13355,13356,13357,13358,13359,13360,13361)) ] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:390]
    * Error in query [SELECT selementid FROM sysmaps_elements WHERE (elementid IN (10057)) AND elementtype=0] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_fetch_assoc() expects parameter 1 to be resource, boolean given[/var/www/html/include/db.inc.php:520]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:463]
    * Error in query [DELETE FROM hosts_groups WHERE (hostid IN (10057)) ] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:463]
    * Error in query [DELETE FROM hosts_templates WHERE (hostid IN (10057)) ] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:390]
    * Error in query [SELECT DISTINCT actionid FROM conditions WHERE conditiontype=1 AND (value IN (10057)) ] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_fetch_assoc() expects parameter 1 to be resource, boolean given[/var/www/html/include/db.inc.php:520]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:463]
    * Error in query [DELETE FROM conditions WHERE conditiontype=1 AND (value IN (10057)) ] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:463]
    * Error in query [DELETE FROM hosts_profiles WHERE (hostid IN (10057)) ] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:463]
    * Error in query [DELETE FROM hosts_profiles_ext WHERE (hostid IN (10057)) ] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:390]
    * Error in query [SELECT DISTINCT ht.* FROM httptest ht, applications ap WHERE (ap.hostid IN (10057)) AND ht.applicationid=ap.applicationid] [ERROR: current transaction is aborted, commands ignored until end of transaction block]
    * pg_fetch_assoc() expects parameter 1 to be resource, boolean given[/var/www/html/include/db.inc.php:520]
    * pg_query() [<a href='function.pg-query'>function.pg-query</a>]: Query failed: ERROR: current transaction is aborted, commands ignored until end of transaction block[/var/www/html/include/db.inc.php:463]
    * Error in query [DELETE FROM hosts WHERE (hostid IN (10057)) ] [ERROR: current transaction is aborted, commands ignored until end of transaction block]

Any idea how to solve this ? Or is there already a newer version of the php frontend available ?
I really prefer postgre above mysql for zabbix, but older versions of postgre are very slow compared to 8.3

Thx
Steven

[teferi]

host deletion is fixed in trunk(I use frontend from there).

btw. some of problems(not all) may be fixed by fine tuning of postgresql