Thread: [Aoetools-discuss] /dev/etherd/err
Brought to you by:
ecashin,
elcapitansam
From: Alex B. <al...@fu...> - 2008-10-08 09:30:34
|
Could somebody please point me at documentation for the /dev/etherd/err? Apart from the aoe documentation included with the linux kernel, which hasn't been very informative about what these errors mean, there hasn't been anything else. I'd be very grateful if someone could point me at any docs or an explanation about what the errors mean, for example: cat /dev/etherd/err retransmit e321.3 oldtag=0017e86a@102aae907 newtag=001ae907 retransmit e321.3 oldtag=0018e86a@102aae907 newtag=001be907 retransmit e321.3 oldtag=0019e86a@102aae907 newtag=001ce907 unexpected rsp e321.3 tag=0017e86a@102aae947 unexpected rsp e321.3 tag=0018e86a@102aae948 unexpected rsp e321.3 tag=0019e86a@102aae94a retransmit e321.1 oldtag=0027e9d7@102aaea2c newtag=002bea2c retransmit e321.1 oldtag=0028e9d7@102aaea2c newtag=002cea2c retransmit e321.1 oldtag=0029e9d7@102aaea2c newtag=002dea2c retransmit e321.1 oldtag=002ae9d7@102aaea2c newtag=002eea2c retransmit e321.3 oldtag=001ce907@102aaea4c newtag=001dea4c unexpected rsp e321.3 tag=001ce907@102aaea85 unexpected rsp e321.1 tag=0027e9d7@102aaea86 unexpected rsp e321.1 tag=0028e9d7@102aaeaa1 unexpected rsp e321.1 tag=0029e9d7@102aaeabe -- Regards, Alex Junior Linux Systems Administrator Fubra Ltd web | www.fubra.com A: Manor Coach House email | al...@fu... : Church Hill phone | +44 (0)844 816 0000 : Aldershot fax | +44 (0)844 816 9999 : Hampshire GU12 4RQ Fubra is a company limited by shares and registered in England and Wales with number 3967214 at Anstey Park House, Anstey Road, Alton, Hampshire, GU34 2RL. We are registered for VAT with number GB733667024, and as a data controller with number Z5193400. We are authorised and regulated by the FSA with number 457679. |
From: Ed C. <ec...@co...> - 2008-10-08 13:47:50
|
On Wed, Oct 08, 2008 at 10:27:49AM +0100, Alex Buell wrote: > Could somebody please point me at documentation for the /dev/etherd/err? That character device is a debugging tool for the aoe driver developers, and although the meaning of the messages in it has stayed relatively stable, their significance has not. Now that there is debugfs, it would be nice if that file could eventually go away, since it was never intended to be a user-level feature. The code in aoechr.c that supports /dev/etherd/err is substantial and tricky enough for me to be repeatedly called on to justify it on the Linux Kernel Mailing List. In any case, please keep in mind that retransmits are not necessarily bad. ... > retransmit e321.3 oldtag=0017e86a@102aae907 newtag=001ae907 This line means that an AoE command to e321.3 with tag 0017e86a had to be retransmitted when the retransmit timer ran at jiffies 102aae907. The tag for the retransmit is 001ae907. ... > unexpected rsp e321.3 tag=0017e86a@102aae947 When a response arrives that is not associated with any outstanding command, it usually means that a command was retransmitted just before the response came in. As the driver tunes in on the right amount of time to wait, this happens sometimes. -- Ed Cashin <ec...@co...> |
From: Alex B. <al...@fu...> - 2008-10-08 16:02:13
|
On Wed, 8 Oct 2008 09:48:17 -0400 Ed Cashin <ec...@co...> wrote: > Now that there is debugfs, it would be nice if that file could > eventually go away, since it was never intended to be a user-level > feature. The code in aoechr.c that supports /dev/etherd/err is > substantial and tricky enough for me to be repeatedly called on to > justify it on the Linux Kernel Mailing List. Oh, that does explain a lot, perhaps when I've time I could work on this and submit patches. We do use aoe a lot in our line of work, and also, I use it at home for backup purposes. [ snip decent explanation ] Thanks, that's explained a lot, but what do the following also mean? retransmit e42.42 oldtag=4ce54511@001d4575 newtag=4cf24575 s=000c296a8c6d d=000c29a52bc9 nout=14 retransmit e42.42 oldtag=4ce84511@001d4575 newtag=4cf34575 s=000c296a8c6d d=000c29a52bc9 nout=14 retransmit e42.42 oldtag=4ceb4511@001d4575 newtag=4cf44575 s=000c296a8c6d d=000c29a52bc9 nout=14 Much appreciated! -- Regards, Alex Junior Linux Systems Administrator Fubra Ltd web | www.fubra.com A: Manor Coach House email | al...@fu... : Church Hill phone | +44 (0)844 816 0000 : Aldershot fax | +44 (0)844 816 9999 : Hampshire GU12 4RQ Fubra is a company limited by shares and registered in England and Wales with number 3967214 at Anstey Park House, Anstey Road, Alton, Hampshire, GU34 2RL. We are registered for VAT with number GB733667024, and as a data controller with number Z5193400. We are authorised and regulated by the FSA with number 457679. |
From: Ed C. <ec...@co...> - 2008-10-08 18:22:59
|
On Wed, Oct 08, 2008 at 05:02:02PM +0100, Alex Buell wrote: ... > Thanks, that's explained a lot, but what do the following also mean? > > retransmit e42.42 oldtag=4ce54511@001d4575 newtag=4cf24575 s=000c296a8c6d d=000c29a52bc9 nout=14 > retransmit e42.42 oldtag=4ce84511@001d4575 newtag=4cf34575 s=000c296a8c6d d=000c29a52bc9 nout=14 > retransmit e42.42 oldtag=4ceb4511@001d4575 newtag=4cf44575 s=000c296a8c6d d=000c29a52bc9 nout=14 There's the source and destination MAC address, along with the number of AoE commands that are currently awaiting a response after being sent. This output is subject to change and isn't really intended for general consumption, so just check the source code for the current meaning if you need to look under the hood. -- Ed Cashin <ec...@co...> |
From: Alex B. <al...@fu...> - 2008-10-08 19:07:09
|
On Wed, 8 Oct 2008 14:22:56 -0400 Ed Cashin <ec...@co...> wrote: > > retransmit e42.42 oldtag=4ce54511@001d4575 newtag=4cf24575 > > s=000c296a8c6d d=000c29a52bc9 nout=14 retransmit e42.42 > > oldtag=4ce84511@001d4575 newtag=4cf34575 s=000c296a8c6d > > d=000c29a52bc9 nout=14 retransmit e42.42 oldtag=4ceb4511@001d4575 > > newtag=4cf44575 s=000c296a8c6d d=000c29a52bc9 nout=14 > > There's the source and destination MAC address, along with the number > of AoE commands that are currently awaiting a response after being > sent. This output is subject to change and isn't really intended for > general consumption, so just check the source code for the current > meaning if you need to look under the hood. So, the nout parameter is the number of outstanding ATA commands needing to be processed? Thanks -- Regards, Alex Junior Linux Systems Administrator Fubra Ltd web | www.fubra.com A: Manor Coach House email | al...@fu... : Church Hill phone | +44 (0)844 816 0000 : Aldershot fax | +44 (0)844 816 9999 : Hampshire GU12 4RQ Fubra is a company limited by shares and registered in England and Wales with number 3967214 at Anstey Park House, Anstey Road, Alton, Hampshire, GU34 2RL. We are registered for VAT with number GB733667024, and as a data controller with number Z5193400. We are authorised and regulated by the FSA with number 457679. |