Wired Intelligent Edge

LONG POST WARNING

So I stumbled upon some interesting readings while troubleshooting a wireless controller (being slow and clients not showing authentication attempts in our Clearpass) issue and wanted to some feedback. I searched the innerweb's, but couldn't find anything so I figured I'd post.

First, sorry for the long post and yes, everything is HPE/Aruba, no 3rd party devices. Also, I haven't opened a support case on this yet, wanted to see what others say or check my basis before proceeding.

Specifically the uplink port between our 6300CX and 6200CX switch that is connected over a J4858C using OM2 fiber

Noticed a bunch of 'TX Drops' when running 'show interface x/x/x statistics human-readable':

Access Switch

--------------------------------------------------------------------------------------------------------------------------------------------
Interface             RX Bytes   RX Pkts  RX Drops  TX Bytes   TX Pkts  TX Drops  RX Bcast  RX Mcast  TX Bcast  TX Mcast  RX Pause  TX Pause
--------------------------------------------------------------------------------------------------------------------------------------------
1/1/28                      2T        2G         0      364G      833M      531K      324M        9M        5M        1M         0         0

The reflexive on the Core Switch:

--------------------------------------------------------------------------------------------------------------------------------------------
Interface             RX Bytes   RX Pkts  RX Drops  TX Bytes   TX Pkts  TX Drops  RX Bcast  RX Mcast  TX Bcast  TX Mcast  RX Pause  TX Pause
--------------------------------------------------------------------------------------------------------------------------------------------
2/1/3                     261G      640M         0        1T        2G         0        4M        1M      324M        9M         0         0

That seemed concerning, so I thought I had a bad fiber which I don't have a method of easily testing other than cleaning the ends end to end and that light feeds through. I cleared the ports and downloaded a 1Gb file and the counters started climbing and was at about 100-200 after the 1Gb file was downloaded.  After looking at a bunch of other switches (JL725A - 6200 24G / JL658A - 6300M SFP) , across other sites, I noticed that all of my 6200's report 'TX DROPS' and the 6300's rarely do. My newer R8Q72A - 6200F 12G are not showing any drops. 

Question 1: Is this correct in that we should see TX Drops on a fiber optic link? Could this be a physical, SFP, switch issue or a bug?

 Then I remember that I can get some light readings on on the transceivers and decided to pull the readings using 'interface x/x/x transceiver detail'.

Access switch above:

Transceiver in 1/1/28
 Interface Name      : 1/1/28
 Type                : 1G-SX / 1G SFP SX
 Transceiver Status  : An HPE pluggable module that is supported in this interface
 Diagnostic Support  : DOM
 Connector Type      : LC
 Wavelength          : 850nm
 Transfer Distance   : 0.00km (SMF), 150m (OM1), 300m (OM2), 0m (OM3)
 Product Number      : J4858C
 Serial Number       : <DELETED>
 Part Number         : 1990-4395

 Status
  Voltage            :  3.4 V
  Power Requested    :  1.0 W
  Power Allocated    :  3.0 W
  Port Max Power     :  3.0 W

  -----------------------------------------------------------
  Lane  Tx Bias           Rx Power          Tx Power         
        (mA)              (mW/dBm)          (mW/dBm)         
  -----------------------------------------------------------
  1     n/a               n/a               0.3157 / -5.01 

 Recent Alarms:

Core Switch from above:

Transceiver in 2/1/3
 Interface Name      : 2/1/3
 Type                : 1G-SX / 1G SFP SX
 Connector Type      : LC
 Transceiver Status  : An HPE pluggable module that is supported in this interface
 Wavelength          : 850nm
 Transfer Distance   : 0.00km (SMF), 150m (OM1), 300m (OM2), 0m (OM3)
 Diagnostic Support  : DOM
 Product Number      : J4858C
 Serial Number       : <DELETED>
 Part Number         : 1990-4395

 Status
  Temperature : 31.0078C

  ------------------------------------------------------------
  Lane  Tx Bias           Rx Power          Tx Power          
        (mA)              (mW/dBm)          (mW/dBm)          
  ------------------------------------------------------------
  1     5.3820            0.2520 / -5.99    n/a               

 Recent Alarms:

So this is where I started to notice another issue. The results from 'interface x/x/x transceiver detail' varied widely. Some report n/a's across all readings, random locations and sometime nothing is printed after the temperature. Some report voltage, others don't. It's really odd. 

Question 2: Is this normal, that the readings are all over the board and not consistent from interface to interface? If that is the case, what is the point of even having this command if it can't be relied upon.

Question 3: If we are to trust the readings for them links that show them, what should be considered a good reading?

Below is an image of the other sites and switches where I grabbed some readings just to show how inconsistent it was. Unless noted all LINK B switches are JL725A - 6200 24G. Site 2 is a SM site and the transceivers are less than 8 months old.

As always, any help, speculation or where I can get more information it would be appreciated.

-Chris

Hi,

How long does the fiber run between the two switches? Has it been tested end-to-end before the transceivers? The J4858C is a 1Gb transceiver specified to transmit up to 1800' over OM2. See below:

I have seen "problems" similar to this:

1). Fiber strength for the run is questionable. Have a professi,onal trouble shoot it and if there, are problems re-terminatere-terminate.

2). Make sure you are using the same revision level of J4858 transceiver, change out the transceivers.

3). Try a different pair if you have more than one pair for the run. (Usually they are teminated with 6 pair.)

Good luck

LONG POST WARNING

So I stumbled upon some interesting readings while troubleshooting a wireless controller (being slow and clients not showing authentication attempts in our Clearpass) issue and wanted to some feedback. I searched the innerweb's, but couldn't find anything so I figured I'd post.

First, sorry for the long post and yes, everything is HPE/Aruba, no 3rd party devices. Also, I haven't opened a support case on this yet, wanted to see what others say or check my basis before proceeding.

Specifically the uplink port between our 6300CX and 6200CX switch that is connected over a J4858C using OM2 fiber

Noticed a bunch of 'TX Drops' when running 'show interface x/x/x statistics human-readable':

Access Switch

--------------------------------------------------------------------------------------------------------------------------------------------
Interface             RX Bytes   RX Pkts  RX Drops  TX Bytes   TX Pkts  TX Drops  RX Bcast  RX Mcast  TX Bcast  TX Mcast  RX Pause  TX Pause
--------------------------------------------------------------------------------------------------------------------------------------------
1/1/28                      2T        2G         0      364G      833M      531K      324M        9M        5M        1M         0         0

The reflexive on the Core Switch:

--------------------------------------------------------------------------------------------------------------------------------------------
Interface             RX Bytes   RX Pkts  RX Drops  TX Bytes   TX Pkts  TX Drops  RX Bcast  RX Mcast  TX Bcast  TX Mcast  RX Pause  TX Pause
--------------------------------------------------------------------------------------------------------------------------------------------
2/1/3                     261G      640M         0        1T        2G         0        4M        1M      324M        9M         0         0

That seemed concerning, so I thought I had a bad fiber which I don't have a method of easily testing other than cleaning the ends end to end and that light feeds through. I cleared the ports and downloaded a 1Gb file and the counters started climbing and was at about 100-200 after the 1Gb file was downloaded.  After looking at a bunch of other switches (JL725A - 6200 24G / JL658A - 6300M SFP) , across other sites, I noticed that all of my 6200's report 'TX DROPS' and the 6300's rarely do. My newer R8Q72A - 6200F 12G are not showing any drops. 

Question 1: Is this correct in that we should see TX Drops on a fiber optic link? Could this be a physical, SFP, switch issue or a bug?

 Then I remember that I can get some light readings on on the transceivers and decided to pull the readings using 'interface x/x/x transceiver detail'.

Access switch above:

Transceiver in 1/1/28
 Interface Name      : 1/1/28
 Type                : 1G-SX / 1G SFP SX
 Transceiver Status  : An HPE pluggable module that is supported in this interface
 Diagnostic Support  : DOM
 Connector Type      : LC
 Wavelength          : 850nm
 Transfer Distance   : 0.00km (SMF), 150m (OM1), 300m (OM2), 0m (OM3)
 Product Number      : J4858C
 Serial Number       : <DELETED>
 Part Number         : 1990-4395

 Status
  Voltage            :  3.4 V
  Power Requested    :  1.0 W
  Power Allocated    :  3.0 W
  Port Max Power     :  3.0 W

  -----------------------------------------------------------
  Lane  Tx Bias           Rx Power          Tx Power         
        (mA)              (mW/dBm)          (mW/dBm)         
  -----------------------------------------------------------
  1     n/a               n/a               0.3157 / -5.01 

 Recent Alarms:

Core Switch from above:

Transceiver in 2/1/3
 Interface Name      : 2/1/3
 Type                : 1G-SX / 1G SFP SX
 Connector Type      : LC
 Transceiver Status  : An HPE pluggable module that is supported in this interface
 Wavelength          : 850nm
 Transfer Distance   : 0.00km (SMF), 150m (OM1), 300m (OM2), 0m (OM3)
 Diagnostic Support  : DOM
 Product Number      : J4858C
 Serial Number       : <DELETED>
 Part Number         : 1990-4395

 Status
  Temperature : 31.0078C

  ------------------------------------------------------------
  Lane  Tx Bias           Rx Power          Tx Power          
        (mA)              (mW/dBm)          (mW/dBm)          
  ------------------------------------------------------------
  1     5.3820            0.2520 / -5.99    n/a               

 Recent Alarms:

So this is where I started to notice another issue. The results from 'interface x/x/x transceiver detail' varied widely. Some report n/a's across all readings, random locations and sometime nothing is printed after the temperature. Some report voltage, others don't. It's really odd. 

Question 2: Is this normal, that the readings are all over the board and not consistent from interface to interface? If that is the case, what is the point of even having this command if it can't be relied upon.

Question 3: If we are to trust the readings for them links that show them, what should be considered a good reading?

Below is an image of the other sites and switches where I grabbed some readings just to show how inconsistent it was. Unless noted all LINK B switches are JL725A - 6200 24G. Site 2 is a SM site and the transceivers are less than 8 months old.

As always, any help, speculation or where I can get more information it would be appreciated.

-Chris

Thanks for replying,

Unfortunately I don't have the OTDR results when it was installed in 2000, but I do know the path of the fiber and it is well under 700' in length. There are other pairs available, which we tried another one with new patch cables and the results are the same. I swapped out the transceivers on both sides and still getting the strange test reads back from the 'transceiver detail' command. 

Based on our other sites readings, I'm leaning towards a bug in the JL725A - 6200 24G read back at least for the 'Tx Drops'. I ran a job last night and checked over 200+ fiber connections in use, almost all of the readings from a JL725A - 6200 24G show high on that reading. Connections between JL658A - 6300M SFP's switches and the R8Q72A - 6200F 12G show no TX Drops. I lean to this conclusion even more since the 'transceiver detail' isn't consistent in anyway.

I ran a long packet test between two endpoints across that fiber last night and they both came back clean after a 12 hours run, but the TX Drops soared from the access switch to the core. I just don't like bad data as it leads me down the wrong path, wasting time and energy, when in reality it is working fine. As to the original cause, forcing the VC out of that wing into our core seems have cleared up the complaints.

When I'm back at the site, I was going to get an independent light reading on the fiber and will schedule this site to have the fiber re-tested with an OTDR, that will be a few months out to get the funding in place. In meantime, I'm going to just ignore those errors unless I'm getting it from the devices that have proven to not be giving false data.

-Chris

Hi,

How long does the fiber run between the two switches? Has it been tested end-to-end before the transceivers? The J4858C is a 1Gb transceiver specified to transmit up to 1800' over OM2. See below:

I have seen "problems" similar to this:

1). Fiber strength for the run is questionable. Have a professi,onal trouble shoot it and if there, are problems re-terminatere-terminate.

2). Make sure you are using the same revision level of J4858 transceiver, change out the transceivers.

3). Try a different pair if you have more than one pair for the run. (Usually they are teminated with 6 pair.)

Good luck

LONG POST WARNING

So I stumbled upon some interesting readings while troubleshooting a wireless controller (being slow and clients not showing authentication attempts in our Clearpass) issue and wanted to some feedback. I searched the innerweb's, but couldn't find anything so I figured I'd post.

First, sorry for the long post and yes, everything is HPE/Aruba, no 3rd party devices. Also, I haven't opened a support case on this yet, wanted to see what others say or check my basis before proceeding.

Specifically the uplink port between our 6300CX and 6200CX switch that is connected over a J4858C using OM2 fiber

Noticed a bunch of 'TX Drops' when running 'show interface x/x/x statistics human-readable':

Access Switch

--------------------------------------------------------------------------------------------------------------------------------------------
Interface             RX Bytes   RX Pkts  RX Drops  TX Bytes   TX Pkts  TX Drops  RX Bcast  RX Mcast  TX Bcast  TX Mcast  RX Pause  TX Pause
--------------------------------------------------------------------------------------------------------------------------------------------
1/1/28                      2T        2G         0      364G      833M      531K      324M        9M        5M        1M         0         0

The reflexive on the Core Switch:

--------------------------------------------------------------------------------------------------------------------------------------------
Interface             RX Bytes   RX Pkts  RX Drops  TX Bytes   TX Pkts  TX Drops  RX Bcast  RX Mcast  TX Bcast  TX Mcast  RX Pause  TX Pause
--------------------------------------------------------------------------------------------------------------------------------------------
2/1/3                     261G      640M         0        1T        2G         0        4M        1M      324M        9M         0         0

That seemed concerning, so I thought I had a bad fiber which I don't have a method of easily testing other than cleaning the ends end to end and that light feeds through. I cleared the ports and downloaded a 1Gb file and the counters started climbing and was at about 100-200 after the 1Gb file was downloaded.  After looking at a bunch of other switches (JL725A - 6200 24G / JL658A - 6300M SFP) , across other sites, I noticed that all of my 6200's report 'TX DROPS' and the 6300's rarely do. My newer R8Q72A - 6200F 12G are not showing any drops. 

Question 1: Is this correct in that we should see TX Drops on a fiber optic link? Could this be a physical, SFP, switch issue or a bug?

 Then I remember that I can get some light readings on on the transceivers and decided to pull the readings using 'interface x/x/x transceiver detail'.

Access switch above:

Transceiver in 1/1/28
 Interface Name      : 1/1/28
 Type                : 1G-SX / 1G SFP SX
 Transceiver Status  : An HPE pluggable module that is supported in this interface
 Diagnostic Support  : DOM
 Connector Type      : LC
 Wavelength          : 850nm
 Transfer Distance   : 0.00km (SMF), 150m (OM1), 300m (OM2), 0m (OM3)
 Product Number      : J4858C
 Serial Number       : <DELETED>
 Part Number         : 1990-4395

 Status
  Voltage            :  3.4 V
  Power Requested    :  1.0 W
  Power Allocated    :  3.0 W
  Port Max Power     :  3.0 W

  -----------------------------------------------------------
  Lane  Tx Bias           Rx Power          Tx Power         
        (mA)              (mW/dBm)          (mW/dBm)         
  -----------------------------------------------------------
  1     n/a               n/a               0.3157 / -5.01 

 Recent Alarms:

Core Switch from above:

Transceiver in 2/1/3
 Interface Name      : 2/1/3
 Type                : 1G-SX / 1G SFP SX
 Connector Type      : LC
 Transceiver Status  : An HPE pluggable module that is supported in this interface
 Wavelength          : 850nm
 Transfer Distance   : 0.00km (SMF), 150m (OM1), 300m (OM2), 0m (OM3)
 Diagnostic Support  : DOM
 Product Number      : J4858C
 Serial Number       : <DELETED>
 Part Number         : 1990-4395

 Status
  Temperature : 31.0078C

  ------------------------------------------------------------
  Lane  Tx Bias           Rx Power          Tx Power          
        (mA)              (mW/dBm)          (mW/dBm)          
  ------------------------------------------------------------
  1     5.3820            0.2520 / -5.99    n/a               

 Recent Alarms:

So this is where I started to notice another issue. The results from 'interface x/x/x transceiver detail' varied widely. Some report n/a's across all readings, random locations and sometime nothing is printed after the temperature. Some report voltage, others don't. It's really odd. 

Question 2: Is this normal, that the readings are all over the board and not consistent from interface to interface? If that is the case, what is the point of even having this command if it can't be relied upon.

Question 3: If we are to trust the readings for them links that show them, what should be considered a good reading?

Below is an image of the other sites and switches where I grabbed some readings just to show how inconsistent it was. Unless noted all LINK B switches are JL725A - 6200 24G. Site 2 is a SM site and the transceivers are less than 8 months old.

As always, any help, speculation or where I can get more information it would be appreciated.

-Chris

Thanks for replying,

Unfortunately I don't have the OTDR results when it was installed in 2000, but I do know the path of the fiber and it is well under 700' in length. There are other pairs available, which we tried another one with new patch cables and the results are the same. I swapped out the transceivers on both sides and still getting the strange test reads back from the 'transceiver detail' command. 

Based on our other sites readings, I'm leaning towards a bug in the JL725A - 6200 24G read back at least for the 'Tx Drops'. I ran a job last night and checked over 200+ fiber connections in use, almost all of the readings from a JL725A - 6200 24G show high on that reading. Connections between JL658A - 6300M SFP's switches and the R8Q72A - 6200F 12G show no TX Drops. I lean to this conclusion even more since the 'transceiver detail' isn't consistent in anyway.

I ran a long packet test between two endpoints across that fiber last night and they both came back clean after a 12 hours run, but the TX Drops soared from the access switch to the core. I just don't like bad data as it leads me down the wrong path, wasting time and energy, when in reality it is working fine. As to the original cause, forcing the VC out of that wing into our core seems have cleared up the complaints.

When I'm back at the site, I was going to get an independent light reading on the fiber and will schedule this site to have the fiber re-tested with an OTDR, that will be a few months out to get the funding in place. In meantime, I'm going to just ignore those errors unless I'm getting it from the devices that have proven to not be giving false data.

-Chris

Hi,

How long does the fiber run between the two switches? Has it been tested end-to-end before the transceivers? The J4858C is a 1Gb transceiver specified to transmit up to 1800' over OM2. See below:

I have seen "problems" similar to this:

1). Fiber strength for the run is questionable. Have a professi,onal trouble shoot it and if there, are problems re-terminatere-terminate.

2). Make sure you are using the same revision level of J4858 transceiver, change out the transceivers.

3). Try a different pair if you have more than one pair for the run. (Usually they are teminated with 6 pair.)

Good luck

LONG POST WARNING

So I stumbled upon some interesting readings while troubleshooting a wireless controller (being slow and clients not showing authentication attempts in our Clearpass) issue and wanted to some feedback. I searched the innerweb's, but couldn't find anything so I figured I'd post.

First, sorry for the long post and yes, everything is HPE/Aruba, no 3rd party devices. Also, I haven't opened a support case on this yet, wanted to see what others say or check my basis before proceeding.

Specifically the uplink port between our 6300CX and 6200CX switch that is connected over a J4858C using OM2 fiber

Noticed a bunch of 'TX Drops' when running 'show interface x/x/x statistics human-readable':

Access Switch

--------------------------------------------------------------------------------------------------------------------------------------------
Interface             RX Bytes   RX Pkts  RX Drops  TX Bytes   TX Pkts  TX Drops  RX Bcast  RX Mcast  TX Bcast  TX Mcast  RX Pause  TX Pause
--------------------------------------------------------------------------------------------------------------------------------------------
1/1/28                      2T        2G         0      364G      833M      531K      324M        9M        5M        1M         0         0

The reflexive on the Core Switch:

--------------------------------------------------------------------------------------------------------------------------------------------
Interface             RX Bytes   RX Pkts  RX Drops  TX Bytes   TX Pkts  TX Drops  RX Bcast  RX Mcast  TX Bcast  TX Mcast  RX Pause  TX Pause
--------------------------------------------------------------------------------------------------------------------------------------------
2/1/3                     261G      640M         0        1T        2G         0        4M        1M      324M        9M         0         0

That seemed concerning, so I thought I had a bad fiber which I don't have a method of easily testing other than cleaning the ends end to end and that light feeds through. I cleared the ports and downloaded a 1Gb file and the counters started climbing and was at about 100-200 after the 1Gb file was downloaded.  After looking at a bunch of other switches (JL725A - 6200 24G / JL658A - 6300M SFP) , across other sites, I noticed that all of my 6200's report 'TX DROPS' and the 6300's rarely do. My newer R8Q72A - 6200F 12G are not showing any drops. 

Question 1: Is this correct in that we should see TX Drops on a fiber optic link? Could this be a physical, SFP, switch issue or a bug?

 Then I remember that I can get some light readings on on the transceivers and decided to pull the readings using 'interface x/x/x transceiver detail'.

Access switch above:

Transceiver in 1/1/28
 Interface Name      : 1/1/28
 Type                : 1G-SX / 1G SFP SX
 Transceiver Status  : An HPE pluggable module that is supported in this interface
 Diagnostic Support  : DOM
 Connector Type      : LC
 Wavelength          : 850nm
 Transfer Distance   : 0.00km (SMF), 150m (OM1), 300m (OM2), 0m (OM3)
 Product Number      : J4858C
 Serial Number       : <DELETED>
 Part Number         : 1990-4395

 Status
  Voltage            :  3.4 V
  Power Requested    :  1.0 W
  Power Allocated    :  3.0 W
  Port Max Power     :  3.0 W

  -----------------------------------------------------------
  Lane  Tx Bias           Rx Power          Tx Power         
        (mA)              (mW/dBm)          (mW/dBm)         
  -----------------------------------------------------------
  1     n/a               n/a               0.3157 / -5.01 

 Recent Alarms:

Core Switch from above:

Transceiver in 2/1/3
 Interface Name      : 2/1/3
 Type                : 1G-SX / 1G SFP SX
 Connector Type      : LC
 Transceiver Status  : An HPE pluggable module that is supported in this interface
 Wavelength          : 850nm
 Transfer Distance   : 0.00km (SMF), 150m (OM1), 300m (OM2), 0m (OM3)
 Diagnostic Support  : DOM
 Product Number      : J4858C
 Serial Number       : <DELETED>
 Part Number         : 1990-4395

 Status
  Temperature : 31.0078C

  ------------------------------------------------------------
  Lane  Tx Bias           Rx Power          Tx Power          
        (mA)              (mW/dBm)          (mW/dBm)          
  ------------------------------------------------------------
  1     5.3820            0.2520 / -5.99    n/a               

 Recent Alarms:

So this is where I started to notice another issue. The results from 'interface x/x/x transceiver detail' varied widely. Some report n/a's across all readings, random locations and sometime nothing is printed after the temperature. Some report voltage, others don't. It's really odd. 

Question 2: Is this normal, that the readings are all over the board and not consistent from interface to interface? If that is the case, what is the point of even having this command if it can't be relied upon.

Question 3: If we are to trust the readings for them links that show them, what should be considered a good reading?

Below is an image of the other sites and switches where I grabbed some readings just to show how inconsistent it was. Unless noted all LINK B switches are JL725A - 6200 24G. Site 2 is a SM site and the transceivers are less than 8 months old.

As always, any help, speculation or where I can get more information it would be appreciated.

-Chris

Michael,

We are planning on some testing next week since school will be out during for spring break. I'll report back the results once I get into the site.

I still haven't opened a support ticket on this topic OR another issue with DAC's reporting the same exact errors across a VSF Stack. It's like opening a can of worms and thinking do I really need to go fishing.

-Chris

Thanks for replying,

Unfortunately I don't have the OTDR results when it was installed in 2000, but I do know the path of the fiber and it is well under 700' in length. There are other pairs available, which we tried another one with new patch cables and the results are the same. I swapped out the transceivers on both sides and still getting the strange test reads back from the 'transceiver detail' command. 

Based on our other sites readings, I'm leaning towards a bug in the JL725A - 6200 24G read back at least for the 'Tx Drops'. I ran a job last night and checked over 200+ fiber connections in use, almost all of the readings from a JL725A - 6200 24G show high on that reading. Connections between JL658A - 6300M SFP's switches and the R8Q72A - 6200F 12G show no TX Drops. I lean to this conclusion even more since the 'transceiver detail' isn't consistent in anyway.

I ran a long packet test between two endpoints across that fiber last night and they both came back clean after a 12 hours run, but the TX Drops soared from the access switch to the core. I just don't like bad data as it leads me down the wrong path, wasting time and energy, when in reality it is working fine. As to the original cause, forcing the VC out of that wing into our core seems have cleared up the complaints.

When I'm back at the site, I was going to get an independent light reading on the fiber and will schedule this site to have the fiber re-tested with an OTDR, that will be a few months out to get the funding in place. In meantime, I'm going to just ignore those errors unless I'm getting it from the devices that have proven to not be giving false data.

-Chris

Hi,

How long does the fiber run between the two switches? Has it been tested end-to-end before the transceivers? The J4858C is a 1Gb transceiver specified to transmit up to 1800' over OM2. See below:

I have seen "problems" similar to this:

1). Fiber strength for the run is questionable. Have a professi,onal trouble shoot it and if there, are problems re-terminatere-terminate.

2). Make sure you are using the same revision level of J4858 transceiver, change out the transceivers.

3). Try a different pair if you have more than one pair for the run. (Usually they are teminated with 6 pair.)

Good luck

LONG POST WARNING

So I stumbled upon some interesting readings while troubleshooting a wireless controller (being slow and clients not showing authentication attempts in our Clearpass) issue and wanted to some feedback. I searched the innerweb's, but couldn't find anything so I figured I'd post.

First, sorry for the long post and yes, everything is HPE/Aruba, no 3rd party devices. Also, I haven't opened a support case on this yet, wanted to see what others say or check my basis before proceeding.

Specifically the uplink port between our 6300CX and 6200CX switch that is connected over a J4858C using OM2 fiber

Noticed a bunch of 'TX Drops' when running 'show interface x/x/x statistics human-readable':

Access Switch

--------------------------------------------------------------------------------------------------------------------------------------------
Interface             RX Bytes   RX Pkts  RX Drops  TX Bytes   TX Pkts  TX Drops  RX Bcast  RX Mcast  TX Bcast  TX Mcast  RX Pause  TX Pause
--------------------------------------------------------------------------------------------------------------------------------------------
1/1/28                      2T        2G         0      364G      833M      531K      324M        9M        5M        1M         0         0

The reflexive on the Core Switch:

--------------------------------------------------------------------------------------------------------------------------------------------
Interface             RX Bytes   RX Pkts  RX Drops  TX Bytes   TX Pkts  TX Drops  RX Bcast  RX Mcast  TX Bcast  TX Mcast  RX Pause  TX Pause
--------------------------------------------------------------------------------------------------------------------------------------------
2/1/3                     261G      640M         0        1T        2G         0        4M        1M      324M        9M         0         0

That seemed concerning, so I thought I had a bad fiber which I don't have a method of easily testing other than cleaning the ends end to end and that light feeds through. I cleared the ports and downloaded a 1Gb file and the counters started climbing and was at about 100-200 after the 1Gb file was downloaded.  After looking at a bunch of other switches (JL725A - 6200 24G / JL658A - 6300M SFP) , across other sites, I noticed that all of my 6200's report 'TX DROPS' and the 6300's rarely do. My newer R8Q72A - 6200F 12G are not showing any drops. 

Question 1: Is this correct in that we should see TX Drops on a fiber optic link? Could this be a physical, SFP, switch issue or a bug?

 Then I remember that I can get some light readings on on the transceivers and decided to pull the readings using 'interface x/x/x transceiver detail'.

Access switch above:

Transceiver in 1/1/28
 Interface Name      : 1/1/28
 Type                : 1G-SX / 1G SFP SX
 Transceiver Status  : An HPE pluggable module that is supported in this interface
 Diagnostic Support  : DOM
 Connector Type      : LC
 Wavelength          : 850nm
 Transfer Distance   : 0.00km (SMF), 150m (OM1), 300m (OM2), 0m (OM3)
 Product Number      : J4858C
 Serial Number       : <DELETED>
 Part Number         : 1990-4395

 Status
  Voltage            :  3.4 V
  Power Requested    :  1.0 W
  Power Allocated    :  3.0 W
  Port Max Power     :  3.0 W

  -----------------------------------------------------------
  Lane  Tx Bias           Rx Power          Tx Power         
        (mA)              (mW/dBm)          (mW/dBm)         
  -----------------------------------------------------------
  1     n/a               n/a               0.3157 / -5.01 

 Recent Alarms:

Core Switch from above:

Transceiver in 2/1/3
 Interface Name      : 2/1/3
 Type                : 1G-SX / 1G SFP SX
 Connector Type      : LC
 Transceiver Status  : An HPE pluggable module that is supported in this interface
 Wavelength          : 850nm
 Transfer Distance   : 0.00km (SMF), 150m (OM1), 300m (OM2), 0m (OM3)
 Diagnostic Support  : DOM
 Product Number      : J4858C
 Serial Number       : <DELETED>
 Part Number         : 1990-4395

 Status
  Temperature : 31.0078C

  ------------------------------------------------------------
  Lane  Tx Bias           Rx Power          Tx Power          
        (mA)              (mW/dBm)          (mW/dBm)          
  ------------------------------------------------------------
  1     5.3820            0.2520 / -5.99    n/a               

 Recent Alarms:

So this is where I started to notice another issue. The results from 'interface x/x/x transceiver detail' varied widely. Some report n/a's across all readings, random locations and sometime nothing is printed after the temperature. Some report voltage, others don't. It's really odd. 

Question 2: Is this normal, that the readings are all over the board and not consistent from interface to interface? If that is the case, what is the point of even having this command if it can't be relied upon.

Question 3: If we are to trust the readings for them links that show them, what should be considered a good reading?

Below is an image of the other sites and switches where I grabbed some readings just to show how inconsistent it was. Unless noted all LINK B switches are JL725A - 6200 24G. Site 2 is a SM site and the transceivers are less than 8 months old.

As always, any help, speculation or where I can get more information it would be appreciated.

-Chris