DR: Needed to send a ping message but failed. 1

Hi All, I am running IDS 11.70 FC7 on AIX server. I am running primary, HDR (async) and RSS server, I have noticed an increase in the frequency of following messages: DR: Needed to send a ping message but failed. 1 I have checked the network connectivity between primary and secondary and it is normal and its not showing any glitches. Can someone help me that what particular i can monitor when it start logging these messages in the online.log, so that i could tune that particular item to improve the performance of both primary & secondary. One thing i noticed that when this message appears in the log, the next checkpoint is taking more time to complete than normal. Also i noticed that during this time, there were some user sessions (in the output of onstat -u command) which were showing the following flags: G-BPX-- (which indicates the waiting for write of the logical log buffer) I have checked the IO activity for the disk and CPU utilizations and they are very normal. Thanks in advance Regards, Khan

can someone help me?

The failing to send a ping rather than the more common ping timeout, seems to mean that the server was for whatever reason unable to start the OS ping command. At least I've always assumed the DR pings were passed out to the OS. Anyway, you may want to check on some of your OS resources like max files, max procs and max user procs (I don't know what AIX calls them). --EEM > -----Original Message----- > From: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] On Behalf Of O > KHAN > Sent: Tuesday, November 12, 2013 4:26 PM > To: ids@iiug.org > Subject: Re: DR: Needed to send a ping message but fail.... [31938] > > can someone help me? > > > *********************************************************************** > ******** > Forum Note: Use "Reply" to post a response in the discussion forum.

Original post: Hi All, I am running IDS 11.70 FC7 on AIX server. I am running primary, HDR (async) and RSS server, I have noticed an increase in the frequency of following messages: DR: Needed to send a ping message but failed. 1 I have checked the network connectivity between primary and secondary and it is normal and its not showing any glitches. Can someone help me that what particular i can monitor when it start logging these messages in the online.log, so that i could tune that particular item to improve the performance of both primary & secondary. One thing i noticed that when this message appears in the log, the next checkpoint is taking more time to complete than normal. Also i noticed that during this time, there were some user sessions (in the output of onstat -u command) which were showing the following flags: G-BPX-- (which indicates the waiting for write of the logical log buffer) I have checked the IO activity for the disk and CPU utilizations and they are very normal. Thanks in advance Regards, Khan Response: Ok, so in newer releases the HDR ping mechanism was changed. Any type of HDR message sent counts as the connection still being ok (since that's all the ping is trying to verify is that data is still getting to the secondary). This new message you are seeing can show up if the dr_prping thread tries to send a ping message, but there are currently no free hdr buffers to put a ping message into (no free hdr buffers when doing async hdr would mean that at the moment your primary is generating hdr buffers faster then the secondary can receive and apply them so it's likely your secondary is falling behind the primary a bit). When the ping thread can't get a hdr buffer (as there is only a small fixed number of them) it sends a message to the secondary to see if it's in a checkpoint (as checkpoints on the secondary can temporarily delay applying incoming log records) and if it is not, it then checks to see if anything has been sent/received since it's last check, and if nothing as been sent/received, it reports that message in the MSGPATH file. I think the code thinks that a longer time would have passed between when we 1st check if anything has been sent/received and the 2nd time we check...in my opinion it's very likely that the time between the 1st and 2nd check is not very large at all and it could easily not have had enough time passed for us to noticed new data sent/received (as we are only looking in second increments). So as long as you are not ping timing out, the message itself is pretty harmless. But we you seeing users waiting on the log buffer it would indicate that either the secondary is at times falling behind the primary, or you are generating logical log data faster then your network will allow it to be sent and ack'd. How fast is the network between your primary and secondary? What is the round trip network travel time (the round trip time will have a significant impact on the number of hdr buffers that can be sent per second between the hdr pair and if you sometimes generate more log data then you will see users start to block waiting to get to the logical log buffers like you show in your post). Jacques Renaut IBM Informix Advanced Support APD Team