Hi Mahesh, Here is the ticket: #1436 MDS (TCP transport) frame gets dropped
Regards, Girish -----Original Message----- From: A V Mahesh [mailto:[email protected]] Sent: Thursday, August 06, 2015 9:45 AM To: [email protected] Subject: Re: [users] MDS receives stale message Hi Girish, Please do mention in ticket about Opensaf Version , any additional patches applied , and any customized configuration of ( dtm.conf ,ect..) , linux system wmem_max/rmem_max values and the purpose of sending huge data using checkpoint service. -AVM On 8/6/2015 9:21 AM, A V Mahesh wrote: > Hi, > > >>1285 is different issue altogether, this is bug in checkpoint service. > > Please raise a #ticket with issue and steps to reproduce ,by > attaching your test Ckpt application. > > -AVM > > > On 8/5/2015 4:53 PM, Girish Nagaraj wrote: >> Hi, >> >> Did not succeed with default buffer size also. >> >> Regards, >> Girish >> >> >> >> -----Original Message----- >> From: Mathivanan Naickan Palanivelu [mailto:[email protected]] >> Sent: Wednesday, August 05, 2015 4:49 PM >> To: [email protected] >> Cc: [email protected] >> Subject: Re: [users] MDS receives stale message >> >> Iam not sure if you are alluding that it could be a flow control >> issue in checkpoint service? >> Also, you have increased the MDS_SOCK_SND_RCV_BUF_SIZE=33554432, >> that's a very big value, and i don't know if there's any benchmark >> data associated with that size! >> Mahesh could comment on this. >> >> A question i have is, did your exercise succeed when using the >> default buffer sizes for MDS_SOCK_SND_RCV_BUF_SIZE and DTM_SOCK_SND_RCV_BUF_SIZE? >> >> >> Thanks, >> Mathi. >> >> ----- [email protected] wrote: >> >>> Hi Mathi, >>> >>> 1285 is different issue altogether, this is bug in checkpoint >>> service. One of the frame gets dropped and rest of the frames are >>> interpreted as stale messages. >>> >>> Regards, >>> Girish >>> >>> -----Original Message----- >>> From: Mathivanan Naickan Palanivelu >>> [mailto:[email protected]] >>> Sent: Wednesday, August 05, 2015 3:13 PM >>> To: [email protected] >>> Cc: [email protected] >>> Subject: Re: [users] MDS receives stale message >>> >>> Hi, >>> >>> Ticket https://sourceforge.net/p/opensaf/tickets/1285/ exists for >>> this issue, now that you have reported it again on 4.6, we shall >>> prioritize this for the next upcoming release. >>> >>> Mathi. >>> >>> >>> ----- [email protected] wrote: >>> >>>> Hi, >>>> >>>> Issue seen in opensaf 4.6 also >>>> >>>> Setup/environment details: >>>> >>>> Linux: Fedora 22, Linux localhost.localdomain 4.0.4-301.fc22.i686 >>> #1 >>>> SMP Thu May 21 13:43:18 UTC 2015 i686 i686 i386 GNU/Linux Opensaf >>>> version: 4.6 [root@localhost sbin]# systemctl --version systemd 219 >>>> +PAM +AUDIT +SELINUX +IMA -APPARMOR +SMACK +SYSVINIT +UTMP >>>> +LIBCRYPTSETUP GCRYPT +GNUTLS +ACL +XZ -LZ4 +SECCOMP +BLKID >>> +ELFUTILS >>>> ++KMOD +IDN >>>> >>>> Also I exported export MDS_SOCK_SND_RCV_BUF_SIZE=33554432 and >>>> set >>>> DTM_SOCK_SND_RCV_BUF_SIZE=33554432 in /etc/opensaf/dtmd.conf >>>> >>>> Modified checkpoint application attached. >>>> >>>> Steps: >>>> 1. start opensaf on active and standby node 2. start ckpt_demo >>>> on active, ./ckpt_demo 1 3. start ckpt_demo on standby, >>>> ./ckpt_demo 0 >>>> >>>> Notice error message in mds.log on standby node >>>> >>>> Please provide fix for this defect. >>>> >>>> Regards, >>>> Girish >>>> >>>> -----Original Message----- >>>> From: Mathivanan Naickan Palanivelu >>> [mailto:[email protected]] >>>> Sent: Thursday, May 07, 2015 7:59 PM >>>> To: [email protected] >>>> Cc: [email protected] >>>> Subject: Re: [users] MDS receives stale message >>>> >>>> Hi, >>>> >>>> There have been some fixes done after 4.3, I think you should try >>> this >>>> on >>>> 4.6 and let us know if the issue is still seen? >>>> >>>> One ticket worth mentioning that you might want to look at related >>> to >>>> buffer size increase and fragmentation handling would be >>>> http://sourceforge.net/p/opensaf/tickets/654/. >>>> >>>> >>>> Thanks, >>>> Mathi. >>>> >>>> ----- [email protected] wrote: >>>> >>>>> Hi, >>>>> >>>>> >>>>> >>>>> We use opensaf 4.3.1 with TCP as transport, in active-standby >>>> setup. >>>>> >>>>> When we push lot of data for checkpointing from active, we get >>>> error >>>>> message in mds.log on standby >>>>> >>>>> >>>>> >>>>> MDTM: Some stale message recd, hence dropping adest= >>>>> >>>>> >>>>> >>>>> Any idea what went wrong? Is there any fix already available for >>>> this >>>>> issue in later releases? >>>>> >>>>> >>>>> >>>>> Regards, >>>>> >>>>> Girish >>>>> >>>>> -- >>>>> . >>>>> >>> -------------------------------------------------------------------- >>> -- >>>>> -------- One dashboard for servers and applications across >>>>> Physical-Virtual-Cloud Widest out-of-the-box monitoring support >>>> with >>>>> 50+ applications Performance metrics, stats and reports that give >>>> you >>>>> Actionable Insights Deep dive visibility with transaction tracing >>>>> using APM Insight. >>>>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y >>>>> _______________________________________________ >>>>> Opensaf-users mailing list >>>>> [email protected] >>>>> https://lists.sourceforge.net/lists/listinfo/opensaf-users >>>> -- >>>> . >>> -- >>> . > > ---------------------------------------------------------------------- > -------- _______________________________________________ > Opensaf-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/opensaf-users -------------------------------------------------------------------------- ---- _______________________________________________ Opensaf-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/opensaf-users -- . ------------------------------------------------------------------------------ _______________________________________________ Opensaf-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/opensaf-users
